Advances in Cancer Research Volume 67

ADVANCES IN CANCER RESEARCH VOLUME 67

This Page Intentionally Left Blank

ADVANCES IN CANCERRESEARCH Edited by

GEORGE F. VANDE WOUDE ABL-Basic Research Program NCI-Frederick Cancer Research and Development Center Frederick, Maryland

GEORGE KLEIN Department of Tumor Biology Karolinska lnstitutet Stockholm, Sweden

Volume 67

ACADEMIC PRESS San Diego New York Boston London Sydney Tokyo Toronto

This book is printed on acid-free paper. @ Copyright 0 1995 by ACADEMIC PRESS. INC All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means. electronic or mechanical, including photocopy, recording, or any information storage and retrieval system. without permission in writing from the publisher.

Academic Press, Inc. A Division of Harcourt Brace & Company 525 B Street. Suite 1900, San Diego. California 92101-4495

United Kingdom Edition published b! Academic Press Limited 24-28 Oval Road. London NW 1 7DX International Standard Serial Number: 0065-230X International Standard Book Number: 0-12-006667-X

PRINTED IN THE UNITED STATES OF AMERICA 95 96 9 7 9 X 99 0 0 B B 9 8 7 6

5

4

3 2 1

CONTENTS

CONTRIBUTORS TO VOLUME67 .......................................

ix

FOUNDATIONS IN CANCER RESEARCH Mutation and Cancer: A Personal Odyssey ALFREDG. KNUDSON,JR. I. 11.

111. IV.

Viruses, Somatic Mutations, and Cancer ............................ Human Cancer Genes ....................... . . . . . . . . . . . . . . . . . . . . . . The Emergence of Antioncogenes ................................. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

7 12 18 19

Oncogenic Role of “Master” Transcription Factors in Human Leukemias and Sarcomas: A Developmental Model A. THOMAS LOOK Introduction ..................................................... Transcriptional Control Genes Altered by Chromosomal Rearrangements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Oncogenic Transription Factors and the Developmental Regulatory Proteins of Drosophila ............................................. IV. Summary and Future Directions ................................... References ....................................................... I. 11.

V

25 26

43 50 52

vi

CONTENTS

Pathways of Chromosome Alteration in Human Epithelial Cancers BERNARD DU~RILLAUX I. I1. I11 . I V. V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ Colorectal Adenocarcinoma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other Epithelial Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microsatellite Instability in Epithelial Tumors . . . . . . . . . . . . . . . . . . . . . . . Concluding R ................ . References . . ..............................................

59

61 68 74

77 78

Genetics of Murine Lung Tumors TOMMASO A . DRAGANI. GIACOMO MANENTI.A N D MARCOA . PIEROTTI I . Historical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Comparative Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Genetic Linkage Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . I\'. Transgenic Models of Lung Tumorigenesis . . . . . . . . . . V. Candidate Lung Tumor Susceptibility Genes . . . . . . . . . . . . . . . . . . . . . . . . VI . Genetics of Lung Tumors, Human . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIl . Conclusions and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

83 84 95 99 100 102 105 106

Molecular Pathogenesis of AIDS-Related Lymphomas GIANLUCA GAIDANO A N D RICCARDO DALLA-FAVERA I. I1. I11 . 1V. V. \'I . VII.

Epidemiology of i\lDS-Related Lymphomas . . . . . . . . . . . . . . . . . . . . . . . . . Clinicopathologic Spectrum of AIDS-Related Lymphomas . . . . . . . . . . . . Natural History of AIDS-Related Lymphomas . . . . . . . . . . . . . . . . . . . . . . . Host Factors Contributing to AIDS-Related Lymphoma Development Role of Viral Infection in AlDS-Related Lymphomagenesis . . . . . . . . . . . Genetic Lesions Involved in AIDS-Related Lymphomas . . . . . . . . . . . . . . Conclusions: Distinct Pathogenetic Pathways in the Development of AIDS-Related Lymphomas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

113 115 118 120 126 132 139 143

vii

CONTENTS

HLA Class I Antigens in Human Tumors

FEDERICOGARRIDO, TERESA CABRERA, MIGUELANGELLOPEZ-NEVOT, AND FRANCISCO RUIZ-CABELLO I. Introduction 11. HLA Class I 111. Role of MHC in T and NK Cell Recognition V. VII.

....... .......

155 157

Alterations of HLA Class I Expression in Human Tumors Conclusions .... References .......................................................

185

Molecular Epidemiology of Epstein-Barr Virus Infection

JAN W. GRATAMA AND INGEMAR ERNBERG I. 11. 111. IV. V. VI.

Introduction ..................................................... EBV Genome and Gene Expression ................................ EBV Typing at the DNA Level (Genotyping) ........................ EBV Typing at the Protein Level (Ebnotyping) ...................... Differential Recognition of EBV Genotypes by the Immune System . . . Final Conclusions: Implications for the Biology of EBV Infection ..... References .......................................................

197 200 204 226 237 241 246

Scatter Factor and Angiogenesis

ELIOTM. ROSENAND ITZHAKD. GOLDBERG Introduction: Scatter Factor (Hepatocyte Growth Factor) and the c-met Receptor ............................................ 11. SF Biologic Actions on Blood Vessel Wall Cells in Vitro and in Vzvo . . . . 111. SF as a Potential Tumor Angiogenesis Factor ........................ IV. Role of SF in Angiogenesis: Hypotheses and Future Directions ....... V. Summary and Conclusions ........................................ References .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I.

257 259 266 272 276 277

...

CONTENTS

Vlll

Role of VEGF-Flt Receptor System in Normal and Tumor Angiogenesis

MASABUMISHIBUYA I. 11. 111. IV.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Structure and Functions of VEGFIVPF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fit Receptor Gene Family: T h e Major Receptors for VEGF . . . . . . . . . . . Regulation of Tumor Growth by the Suppression of VEGF-Fit Receptor System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

281 283 296 309

311

317

CONTRIBUTORS Numbers in parentheses indicate the pages on which the authors' contributions begin.

'TERESACABRERA, Seruicio de Ancilisis Chicos e Inmunologia, Hospital Virgen de las Nieves, Universidad de Granada, 18014 Granada, Spain (155) RICCARDODALLA-FAVERA, Division of Oncology, Department of Pathology, College of Physicians &' Surgeons of Columbia University, New York, New York I0032 ( 1 13) TOMMASO A. DRAGANI, Division of Experimental Oncology A, Istituto Nazionale Tumori, 20133 Milan, Italy (83) BERNARD DUTRILLAUX, URA 620, CNRS-Institut Curie, 75231 Paris, France (59) INGEMAR ERNBERG, Microbiology and Tumorbiology Center, Karolinska Institute, S-171 77 Stockholm, Sweden (197) GIANLUCA GAIDANO, Laboratorio di Medicina e O n c o l o p Molecolare, Dipartimento di Scienze Biomediche e Oncologaa Umana, Universita di Torino, Ospedale San Luigz Gonzaga, Turin, Italy (1 13) FEDERICO GARRIDO, Seruicio de Andisis Clinicos e Inmunologia, Hospital Virgen de las Nieves, Universidad de Granada, I8014 Granada, Spain (155) ITZHAK D. GOLDBERG, Department of Radiation Oncology, Long Island Jewish Medical Center, The Long Island Campus for Albert Einstein College of Medicine, New Hyde Park, New York 11042 (257) JANW. GRATAMA, Department of Clinical and Tumor Immunology, Daniel den Hoed Cancer Center, Rotterdam, The Netherlands (197) ALFRED G. KNUDSON, JR., Fox Chase Cancer Center, Institute for Cancer Research, Philadelphia, Pennsylvania I 9 1 I I ( 1) A. THOMAS LOOK,Department of Experimental Oncology, St. Jude Children's Research Hospital and University of Tennessee College of Medicine, Memphis, Tennessee 38105 (25) MIGUELANGELLOPEZ-NEVOT, Semicio de Andisis Clinicos e Inmunologia, Hospital Virgen de lus Nieves, Universidad de Granada, I8014 Granada, Spain (155) GIACOMO MANENTI,Division of Experimental Oncology A, Istituto Nazionale Tumori, 20133 Milan, Italy (83) ix

T

MARCOA. PIEROTTI, Daimion of Expenmental Oncology A , Istituto Nazaonale Tumorz, 20133 Malan, Italy (83) ELioT M. ROSEN, Department of Radaation Oncology, Long Island Jewzsh Medacal Center, The Long Island Campiu for Albert Eanstezn College of Medicane, “Vew Hyde Park, New York 11042 (257) FRANCISCORUIZ-CABELLO, S e n i m o de Ancilsu Clinacos e Inmunologia, Hospital Virgen de las Nzezies, Unzuerszdad de Granada, 18014 Granada, Spuan (155) MASABUMISHIBUYA, Institute of Medical Science, Unzversity of Tokyo, Tokyo 108, Japan (281)

FOUNDATIONS IN CANCER RESEARCH MUTATION AND CANCER: A PERSONAL ODYSSEY Alfred G. Knudson, Jr. Fox Chase Cancer Center, Institute for Cancer Research, Philadelphia, Pennsylvania 19111

I. Viruses, Somatic Mutations, and Cancer A. An Introduction to Cancer in Children B. Oncogenic Viruses C. Somatic Mutations 11. Human Cancer Genes A. Hereditary Cancer and Two Hits B. Oncogenes and Protooncogenes 111. The Emergence of Antiocogenes A. RBI, the First Human Antioncogene B. RBI, TP53,and DNA Tumor Viruses C. Genetic Events in Carcinogenesis IV. Conclusions References

1. Viruses, Somatic Mutations, and Cancer

A. AN INTRODUCTION TO CANCER IN CHILDREN It seems impossible that a student of 50 years ago could live to witness the assembly of our present state of knowledge about cancer, imperfect though it may be. Imagine a time when no student was taught that genes were made of DNA, and no child had ever been cured of leukemia. All of us who have experienced the intervening developments have had different vantage points and different foci of interest. Peter Nowell has provided a superb account of much of this time from the point of view of a cytogeneticist, with an excellent background of earlier events, including the seminal ideas of Boveri on the somatic mutational origin of cancer (Boveri, 1914; Nowell, 1993). My own interest in cancer is rooted in genetics and pediatrics. My enthusiasm for genetics dates back to my undergraduate experience at the California Institute of Technology in 1942 with Professors Arthur Sturtevant and Thomas Hunt Morgan; my curiosity about cancer arose during a residency in pediatrics at New York Hospital that included a rotation in 1949 at Memorial Sloan Kettering Cancer Center. It was Dr. 1 ADVANCES IN CANCER RESEARCH, VOL. 67

Copyright 0 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

2

ALFRED G. KNUDSON, J R .

Harold Dargeon, later the author of a well-known book on cancer in children (Dargeon, 1960), who introduced me to that subject. Also there were two pioneers of cancer treatment, Drs. David Karnofsky and Joseph Burchenal, who were administering folic acid antimetabolites to produce some of the first remissions in acute lymphocytic leukemia. However, as exciting as this new development was, even more so to me were the questions: How do young children, even newborn babies, acquire cancer! Why is much of this cancer unique to children? This interest in cancer lay fallow for several years, through pediatric training, Arm); service, and graduate education, until my first position as chairman of a small pediatrics department at the City of Hope Medical Center, where I was responsible for the care o f children with cancer. In 1956 the excitement that followed the first induction of remissions in children with leukemia was still in the air, but by then there was also the realization that most of these children eventually relapsed and died. I became more and more interested in the origin of cancer in children, especially in the cause of leukemia. A survey of 108 cases gave only a few clues, of the kinds already known (Knudson, 1965a). Thus, for 17 there was a significant history of radiation exposure, either in utero or after birth, 3 had a predisposing condition, Down’s syndrome, and 1 had a family history of leukemia, but for the rest there was no clue. It seemed that most cases had arisen from chance events, such as somatic mutations or rare malignant transformation by an undetected latent virus. €3. ONC,OGENIC VIRUSES

In preparing a book, Genetzcs and Dueme (Knudson, 1965b), that included a chapter on cancer, I became particularly interested in the viral possibility and undertook some work with Marcel BaIuda, who had also moved to the City of Hope from Caltech in 1956, where he had been a graduate student in Renato Dulbecco’s pioneering tumor virus group. After his move, Baluda began studying avian myeloblastosis virus ( A M V ) , whose effects were especially interesting to me because they included not only acute myeloid leukemia but also lymphomatosis, osteogenic sarcoma, and nephroblastoma, which closely resembled Wilms’ tumor in children. (Later it was shown that these last three disorders were caused by a myeloblastosis-associated virus.) With infection of chicks before hatching, Baluda was able to develop a sensitive in uiuo assay for the detection of the defective AMV and its helper virus (Baluda and Jamieson, 1961). He also developed an zn zutro assay for the inlestigation of transformation of AMV (Baluda and Goetz, 1961). It seemed possible then that a \irus could also account for human

MUTATION AND CANCER: A PERSONAL ODYSSEY

3

leukemia, because occasional clusters of cases in time and place were reported, and because some RNA viruses only occasionally caused cancer. Our efforts to isolate a tumor virus from human leukemic cells were fruitless, but they stimulated my desire to work in Baluda’s lab on AMV. Howard Temin, who had done his graduate work with Harry Rubin in Dulbecco’s lab, as well as John Bader had shown that certain inhibitors of DNA synthesis interfered with the replication of Rous sarcoma virus (RSV) (Temin, 1963, 1964a; Bader, 1964), so Baluda wanted to test whether that was an aberration or a more general feature of RNA tumor viruses. Inhibition proved to be true for AMV as well (Knudson et al., 1967). T h e most intriguing explanation was that DNA was an obligate intermediate in the replication of AMV, as already suggested by Temin, amd that this viral DNA was integrated into the host genome. Furthermore, there must be an enzyme that can catalyze the synthesis of DNA from RNA. The latter enzyme, reverse transcriptase, was discovered by Temin and David Baltimore in 1970 (Temin and Mizutani, 1970; Baltimore, 1970). Meanwhile, Dulbecco and others were investigating polyoma virus and other DNA tumor viruses. Polyoma virus could transform hamster cells in vitro (Vogt and Dulbecco, 1960) and produced a tumor antigen in them (Habel, 1962). Fried (1965), in Dulbecco’s group, found a temperature-sensitive mutant of polyoma upon which transformation depended. Two other DNA viruses, simian virus 40 (SV40) and certain strains of adenoviruses, also produced such antigens (Black et al., 1963; lluebner et al., 1963; Pope and Rowe, 1964). Although the cells revealed these antigens, they did not produce virus. The production of chromosome abnormalities in transformed cells suggested that virus was still present, probably integrated into host DNA (Shein and Enders, 1962; Koprowski et al., 1962; Vogt and Dulbecco, 1963). This idea was subsequently proved correct (Sambrook et al., 1968; Hirai and Defendi, 1972; Burger and Doerfler, 1974). The antigens later were shown to be specified by viral genes and responsible for transformation: T. antigen for SV40 (Brugge and Butel, 1975; Martin and Chou, 19’75 Osborn and Weber, 1975; Tegtmeyer, 1975) and E1A and E1B for certain adenoviruses (Graham et al., 19’75; Sharp et al., 1975). Specific genes also accounted for transformation by the RNA tumor viruses. Both RSV and AMV were acutely transforming, although some RNA viruses did not cause cancer at all or only after a latent period. What was the difference between these kinds of viruses? A specific transforming gene was suggested by temperature-sensitive mutants of RSV that transformed cells at permissive temperatures but not at nonpermissive temperatures (Toyoshima and Vogt, 1969; Martin, 1970). The ge-

4

ALFRED G . KNUDSON, JR.

nomes of acutely transforming viruses contained some RNA not present in the slowly transforming ones (Duesberg and Vogt, 1970); the part missing in the latter corresponded to the location of the temperaturesensitive gene of RSV (Wang et al., 1975). This was indeed a transforming gene, which came to be known as the STC oncogene for its transformation of fibroblasts into sarcoma cells. For both the RNA and DNA tumor viruses, the key questions became: What are the transforming genes? How d o they work? T h e acutely transforming viruses rarely occur in nature, arising de novo from slow viruses. T h e RNA viruses that are slowly oncogenic can infect large numbers of animals. Thus, a flock of chickens or a colony of mice may carry such viruses and even transmit them from one generation to another. They often succumb to virus-induced cancer, e.g., lymphomatosis in chickens and lymphoma in mice. Since some animals die before the reproductive period ends, there would be natural selection against the causative viruses, unless there were some balancing “good” effect. For example, such viruses might interfere with infection by certain virulent viruses, providing a selective advantage that could maintain the latent virus in a population. An equilibrium state could ensue that over many generations might foster “integration of a latent virus into the host cell, and even into its genetic machinery, . , . complicating discrimination between virus as parasite and virus as organelle or as part of the genome” (Knudson, 1966). Huebner and Todaro (1969) proposed that integrated RNA viruses could provide a critical transforming gene, which they called an oncogene. Such an oncogene could be responsible not only for transformation by acutely transforming viruses but also, in its integrated form, be a target for somatic mutations, without external intervention of virus. Indeed, the viral and somatic mutation hypotheses on the origin of cancer might not be so different after all.

C. SOMATIC MUTATIONS

1. Evidence for Mutation The somatic mutation hypothesis languished for years for lack of support. T h e first such support came from the discovery by Muller that ionizing radiation, already known to be carcinogenic, was also mutagenic (Muller, 1927). When they discovered that radiation could cause deamination of cytosine in DNA, Ponnamperuma et al. (1962) pointed out that cytosine would be changed to uracil, in turn converting a C-G pair to an A-T pair. Evidence for the induction of cancer by ultraviolet lightinduced mutation was provided by the discovery that xeroderma pig-

MUTATION AND CANCER: A PERSONAL ODYSSEY

5

mentosum, a condition that predisposes one to sunlight-induced cancer of the skin, is caused by a defect in the repair of UV-induced thymine climers in DNA (Cleaver, 1968). The discovery that some carcinogenic chemicals were also mutagenic further supported the hypothesis. In the 1960s, certain chemical carcinogens were shown to interact with, and alter, DNA (Brookes and Lawley, 1964). Unfortunately for the somatic mutation hypothesis, some carcinogens were not found to be mutagenic in bacterial test systems, an anomaly that remained for Bruce Ames, who had also been a graduate student at Caltech in the 1950s, to solve later (Ames et al., 1973). Knowing that some carcinogenic compounds must be activated from a procarcinogenic state (Miller et al., 1961), Ames used mammalian tissue extracts that could activate carcinogens that were not activated by bacteria. Boveri’s hypothesis had been based upon reports of mitotic abnormalities in cancer, but of course those could be secondary rather than primary events. Support for the primacy of genetic change came in 1960 with the stunning discovery by Peter Nowell and David Hungerford of a specific karyotypic abnormality in a particular cancer, viz., the Philadelphia chromosome (Phl) in chronic myelocytic leukemia (CML) (Nowell and Hungerford, 1960). The abnormality, which seemed to involve the loss of part of chromosome 22, was found in nearly all cases and the karyotypes were otherwise normal. This finding, and evidence from ><-inactivationstudies that cancers, including CML (Fialkow et al., 1967), were derived from single cells, as originally proposed by Boveri, sudclenly gave credence to the somatic mutation hypothesis. Such a mutation could result from a spontaneous genetic event or from the action of agents like chemical carcinogens, ionizing radiation, or tumor viruses; i.e., it could be a final common pathway that could unite the disparate ildeas of carcinogenesis. T h e question then became one of identifying the genetic changes that could mediate carcinogenesis. 2 . How Many Events?

By 1970 other threads had been woven into the fabric of the picture of cancer. One group of investigators was attempting to understand the strong dependence of the incidence of human cancer upon age. I n the 1950s, Nordling (1953) and Armitage and Doll (1954) observed that the incidence of some cancers rose with approximately the 6th power of age and suggested that the phenomenon might be explained by multiple independent mutational events. Assuming that cancers arise from single cells, they reasoned that some critical number of events in one cell would be necessary to convert it to a cancer cell. Age-specific incidence (I)and age (t) would be expected to bear the relationship, I = k - 1 , where r is

6

ALFRED G . KNUDSON, JR.

the number of events. Therefore, if I = kt6, r = 7. Some important assumptions in such a mathematical model include the constancy of mutation rate with age, constancy of the target cell population with age, and lack of growth advantage of intermediate cells. Failure to make an allowance for the growth advantage of intermediate cells would lead to an overestimate of the required number of events. In a later model, Armitage and Doll (1957) proposed that the number of events could be as few as two if the first provided sufficient growth advantage. The idea that cancer might occur in stages was already known for some experimental situations. Thus, in their classical experiments on chemical carcinogenesis, Berenblum (Berenblum, 1941; Berenblum and Shubik, 1947) and Rous (Rous and Kidd, 1941; Friedewald and Rous, 1944) showed that small doses of only carcinogens on the skin of animals may produce no cancers, but, if carcinogen is followed by croton oil or wounding, cancers will appear. Rous and Kidd (194 1) referred to these two processes as initiation and promotion. The former was considered to involve mutation. Promotion evidently stimulated replication of target stem cells, so that many cells would then carry a mutation induced in a single cell by an initiating carcinogen. The idea that cancer could be produced in a single step was not viable, because of the well-known latent period between exposure to an environmental carcinogen and the appearance of cancer, and because the dose-response curves for such carcinogens were not linear. Furthermore, tumor growth was associated with the evolution of malignancy, sometimes abruptly, in a general process referred to as progression (Foulds, 1954). Although some viruses might produce cancer in a single step, other viruses, and chemical carcinogens, evidently required more. That the target cell number could vary with time was apparent just from a consideration of pediatric cancers. The typical ones are “pediatric” because they have age-specific incidence peaks in childhood, reflecting the disappearance with age of the relevant target cells, e.g., retinoblast, neuroblast, and nephroblast. T h e previously noted models therefore could not account for these cancers; an entirely different mathematical model became necessary (Hethcote and Knudson, 1978). Since some cases of retinoblastoma, neuroblastoma, and Wilms’ tumor are discovered at birth, it is also apparent that the critical number of causative events may be small. Thus, added to the question of the nature and number of carcinogenic events is the question of their targets. A two-stage mathematical model, constructed for cancer generally, took into consideration not only the growth of target cells, but also the difference between the birth rate and death rate of intermediate cells (Mool-

MUTATION AND

CANCER: A

PERSONAL ODYSSEY

7

gavkar and Knudson, 1981). This model could be fitted very well to incidence data on common cancers.

II. Human Cancer Genes

A. HEREDITARY CANCER AND Two HITS When R. Lee Clark invited me in 1969 to move to the M. D. Anderson Cancer Center in Houston in order to establish a medical genetics center there, I turned to the phenomenon of the inheritance of susceptibility to cancer. T h e tumor viruses and the hereditary cancers both seemed to hold the promise of identifying specific genetic factors in carcinogenesis; radiation and chemicals did not d o so. I had seen patients with neurofibromatosis, hereditary retinoblastoma, and familial leukemia, but had never studied the problems. I had had one patient with neurofibromatosis whose parents were both normal; she obviously had a new mutant gene. She had neuroblastoma as an infant, suggesting that perhaps some cases of that disease are due to new germ line mutation, so that the disease might represent a mixture of hereditary and nonhereditary forms both caused by mutation, whether germinal or somatic or both. We proposed that germ line mutations for neuroblastoma would usually be lethal (dominant lethal) and that those children with more than one neuroblastoma would be especially good prospects for germ line mutation (Knudson and Amromin, 1966). Neuroblastoma might resemble retinoblastoma, and so too might other cancers with a high mortality rate before the age of reproduction. Retinoblastoma was especially interesting for further investigation. T h e very significant cure rate permitted survival until the age of reproduction in many cases. In fact, hereditary cases were well-known. Furthermore, about 25-30% of cases were affected bilaterally. These child r e n generally had normal parents, but half of the offspring of such survivors were affected, indicating that for all practical purposes all bilateral cases represent germ line mutations (Schappert-Kimmijser et al., 1966). Falls and Nee1 (1951) had proposed that the sporadic, nonhereditary cases resulted from mutation at the same locus that was mutant in the germ line cases. However, it was recognized that a few individuals who were obligate carriers of the mutation did not develop the tumor, i.e., penetrance was incomplete, even though very high (approximately 95%). Others developed only unilateral tumors. It was obvious that inheritance of the retinoblastoma mutation was not a sufficient condition for tumorigenesis; some other event was necessary. In fact, at

8

ALFRED G . KNUDSON, JR.

the cellular level it must be a rare event, since every retinoblast would carry the mutation in the germ line cases. If a transforming event occurred only a few times in carriers of the germ line mutatioc, then one might expect that some carriers would develop no tumor, some, unilateral tumors, and some, bilateral tumors. In fact, the distribution of the number of tumors approached a Poisson mean of three tumors per carrier (Knudson, 1971). Three retinoblasts among the millions that exist at some time in each eye are transformed into tumor cells. This frequency immediately suggested a somatic mutation rate. One could then hypothesize that tumors in carriers result from two mutations: the first in the germ line the second in somatic retinoblasts. T h e nonhereditary cases could then be explained by two somatic mutations in the same cell. This explanation was supported by an analysis of the ages at which the patients developed tumors; the hereditary cases fit a one somatic hit (mutation) curve and the nonhereditary cases a two somatic hit curve. Both forms would be rare, because the hereditary cases (2 per 100,000 births) are sustained in a population by germ line mutations, and the nonhereditary cases (3 per 100,000 births) require two somatic mutations in the same cell. An estimate of the somatic mutation rate in a germ line case was in close accord with estimates of the two somatic mutation rates in a nonhereditary case (Knudson, 197 1 ; Knudson et al., 1975; Hethcote and Knudson, 1978). The mathematical model that Herbert Hethcote and I developed took into account the growth, differentiation, and ultimate disappearance of the target cells (Hethcote and Knudson, 1978). Louise Strong and 1 took the same approach in the analysis of neuroblastoma and Wilms’ tumor, two other embryonal tumors of children that occur in paired organs (adrenal medulla and kidney), which are also found in both hereditary and nonhereditary forms. The frequencies of the familial forms are much lower, even though the incidences of these tumors are significantly greater than that of retinoblastoma. However, if one assumes that bilateral cases without family history nevertheless are associated with germ line mutations, most of them new, then a higher fraction of cases is attributable to a carrier state, just as with retinoblastoma. We therefore concluded that the two mutation hypothesis was applicable to all three of these embryonal tumors (Knudson and Strong, 1972a,b). I was still haunted by the thought of insertion of a viral genome into the host genome, however, and the report of vertical transmission of nephroblastoma-inducing virus in chickens (Lacour et al., 1970) only promoted that possibility (Knudson and Strong, 1972b).

MUTATION A N D CANCER: A PERSONAL ODYSSEY

9

If two mutations are responsible for some cancers, do they occur in two different genes or in the two copies of a single gene? There was no compelling evidence in 1971. Lele et al. (1963) were the first to report a case of retinoblastoma in which there was a constitutional deletion in a D group chromosome (numbers 13- 15). Such deletions were associated with congenital defects, notably mental retardation. Presumably, one copy of a gene predisposing to retinoblastoma was deleted. For Wilms’ tumor there were rare cases that were associated with aniridia, mental retardation, and genitourinary abnormalities (Miller et al., 1964). We suggested that they could be attributed to the deletion of more than one gene from some chromosome, one of which was a Wilms’ tumor gene (Knudson and Strong, 1972b). This later proved to be the case (Ladda et al., 1974; Riccardi et al., 1978). In no instance was there a deletion of both chromosomes, but theq that very possibly would be lethal to the cell. If the second copy of the tumor gene were defective, it would almost necessarily be a subtle, submicroscopicchange, perhaps limited to the gene itself. If so, the gene would be recessive for oncogenesis, the dominant designation being suitable only for description of transmission of the predisposing state. The idea that recessive cancer genes might exist was originally proposed by Boveri. In the 1960s several investigators used the fusion of cancer cells with normal cells to ascertain whether the product would be normal or malignant. The first investigators found that it was malignant, but Henry Harris and his colleagues (Harris et al., 1969) discovered that it was not malignant until it lost chromosomes. They surmised that such tumor suppression reflected the existence of one or more recessive genes in the normal cells that were lost or mutated in the malignant cells. ‘The relevance of suppression for human cancer was further substantiated by Eric Stanbridge (1976). In 1971 there was not much basis for speculation about the nature of a recessive cancer gene. My own favorite candidate was a gene that coded for a cell surface tissue-specific recognition molecule. In considering hereditary cancers, it was apparent that susceptibility was never to cancer generally; it was typically to one or a few specificcancers. Since it was lknown that tissues displayed tissue-specific recognition molecules on their surfaces, one could imagine that the loss of both copies of such genes could lead to failure of self-recognitionand to cell division (Knud:Son, 1973). On the other hand, David Comings (1973) proposed that the .retinoblastoma gene might belong to a class of genes that directly regulate the activity of genes that stimulate cell division. Only by discovering ithe genes themselves could their nature be ascertained.

10

ALFRED G. KNUDSON,

JR.

A N D PROTOONCOGENES B. ONCOGENES

New developments in cytogenetics helped in movement toward this goal of gene discovery, but for a different kind of cancer gene. New techniques for banding chromosomes had a great impact on the genetics of cancer, as reviewed by Nowell (1993). First, there were new discoveries of tumor-specific changes. Second, chromosomal translocation was found to underlie two important conditions, Burkitt’s lymphoma and CML. In CML the Phl proved to involve a reciprocal translocation between chromosomes 9 and 22, notjust a deletion of 22 (Rowley, 1973). In Burkitt’s lymphoma there was an abnormal chromosome 14 (Manolov and Manolova, 1972), resulting from a reciprocal translocation between chromosomes 8 and 14 (Zech et al., 1976; Manolova et al., 1979). Third, chromosomal deletions associated with cancer could be defined with such precision that comparison of deletion cases could define a common band of deletion, to which a suspected cancer gene could be localized. Thus, the retinoblastoma gene was localized to chromosomal band 13q14 (Francke and Kung, 1976; Knudson et al., 1976). These findings clearly showed that there were two fundamentally different kinds of genetic change associated with cancer, translocation and deletion. T h e deletions pointed to the importance of genetic losses and perhaps recessive genes, but the translocations remained puzzling. This puzzle would not be solved without new discoveries in molecular genetics. These discoveries began with a new understanding of RNA tumor viruses, by then known as retroviruses because of their reverse transcriptase activity. DNA hybridization studies indicated that DNA from normal chicken cells could hybridize with DNA that is complementary to the RNA of the avian tumor viruses, i.e., some part of the viral genome was homologous to a part of the host genome (Temin, 1964b; Baluda and Nayak, 1970; Rosenthal et al., 1971). Some of this DNA represented whole viral genomes. However, it was subsequently discovered by Stehelin, Varmus, Bishop, and Vogt (Stehelin et al., 1976) that the src gene of RSV was like part of a normal cellular gene. Indeed, for the oncogenes that render certain viruses acutely transforming, there is a parallel series of homologous host genes. T h e acutely transforming viruses were created by recombination between host DNA and the DNA of a slow virus. Such an oncogene is a truncated host protooncogene. The viruses can activate transcription of these oncogenes to levels much beyond those normally found for the expression of the host protooncogene. It was then discovered that retroviruses lacking such oncogenes could on occasion insert complementary DNA adjacent to a host protooncogene, thereby activating the latter (Hayward et al., 1981). Such viruses must be

MUTATION AND CANCER: A PERSONAL ODYSSEY

11

inserted at a particular locus, whereas the acutely transforming viruses insert randomly. Long terminal repeat (LTR) DNA of the integrated virus normally acts as a promoter of the transcription of viral genes, but can also activate an adjacent host gene by the process of promoter insertion. Host protooncogenes could be agents of cancer when activated, thereby raising a new question: Could such genes be activated by other than viral means, and so constitute a class of dominantly acting cancer genes distinct from a class of recessively acting cancer genes? T h e answer was to be provided by Burkitt’s lymphoma. Although an 8; 14 translocation characterizes most cases of Burkitt’s lymphoma, some variants were found with 2;8 or 8;22 translocations. Then chromosomes 14, 2, and 22 were identified as the carriers of the hieavy chain, K chain, and A chain immunoglobulin genes, respectively, and chromosome 8 was identified as the site of the c-my protooncogene, tlhe homologue of the oncogene of the myelocytomatosis virus. The promoter insertion model was quickly adapted to this situation, with the hLypothesis that promoting immunoglobulin sequences were activating the protooncogene as the result of juxtaposition by translocation (Klein, 1981). This proved to be the case (Dalla-Favera et al., 1982a; Taub et al., 1982; Erikson et al., 1983; Hamlyn and Rabbitts, 1983). Since then some other translocations have been shown to produce this result in lymphoid neoplasms, although the more usual translocations, those associated with all nonlymphocytic leukemias, some lymphomas and lymphocytic leukemias, and sarcomas, produce fusion genes and proteins, as in the c,ase of CML. Thus for CML, one of the genes involved in the fusion is homologous to the oncogene of the Abelson retrovirus (DeKlein et al., 1982). In most cases the fusion gene proteins are novel transcription factors that act as dominant gain of function mutations [reviewed by Rabbitts (1994)l. About the same time another kind of protooncogene activation was being demonstrated. T h e approach here was to ascertain whether a dominantly acting transforming gene could be transfected from a tumor to a normal cell and transform it. In fact, Robert Weinberg and his colleagues accomplished this (Shih et al., 1979). The offending oncogene in such transformations proved to be homologous to a known sarcoma gene of the ras family (Der et al., 1982; Parada et al., 1982; Santos et al., 1982), originally discovered in certain rat sarcoma viruses. The transforming genes were later shown to differ from the normal ones by single base mutations. The exciting phenomenon of in vitro transformation by oncogenes was then studied intensively. The original target cells were immortalized, but not transformed. When normal cells, not immortalized, were used as

12

ALFRED G. KNUDSON, .JR.

targets, transformation was more difficult to achieve, and indeed Weinberg and his colleagues found that two different oncogenes were required (Land et al., 1983). They fell into two groups: the members of one being substitutes for rus and those of the other being substitutes for my; for example, the adenoviral ElA gene and rus cooperated in transformation (Ruley, 1983). Oncogenes also appeared on the scene with amplification of c-my in a myeloid leukemia line (Collins and Groudine, 1982; Dalla-Favera et al., 1982b). Cytogeneticists had long known about double minute (DM) chromosomes, acentric fragments of chromosomes typically occurring in multiple copies in chromosome preparations. Closely related are intrachromosomal segments that d o not show banding, but interrupt the normal banding pattern in a chromosomal region. These homogeneously staining regions (HSRs) and DMs were shown for the cells of a mouse tumor to contain 30-60 copies of the N-myc protooncogene (Schwab et al., 1983). T h e m y family has been the most frequently amplified kind of gene, although others have been found. T h e kinds of human tumors that show oncogene abnormalities are not random. As noted earlier, tumor-specific translocations curiously are limited to certain hematopoietic neoplasms and sarcomas, e.g., Ewing’s sarcoma and alveolar rhabdomyosarcoma. On the other hand, ras mutations are found in a significant fraction of many kinds of carcinomas. Amplified oncogenes occur in a variety of tumors, including embryonal tumors and carcinomas, but usually not in tumors with specific translocations. Only the specific translocations seem to be primary changes that initiate selected neoplasms; the others are secondary. Most kinds of cancer seem to be initiated by other mechanisms.

ill. The Emergence of Antioncogenes A. R B I ,

THE

FIRSTHUMAN ANTIONCOGENE

In December of 1976 I moved to Philadelphia to become Director of The Institute for Cancer Research at the Fox Chase Cancer Center. At that very time, Baruch Blumberg was in Sweden receiving the Nobel Prize for his discovery of the hepatitis B virus, which was later shown by Jesse Summers and William Mason (1982) to belong to a unique class of DNA viruses with reverse transcriptase activity, and whose mechanism of hepatocarcinogenesis still has not been elaborated. By then it was becoming clear that retinoblastoma was a prototypic tumor that provided the best opportunity to discover the first example of a human recessive cancer gene. It was postulated that the second event could be a

MUTATION A N D CANCER: A PERSONAL ODYSSEY

13

mutation within the gene, deletion, chromosomal loss, or mitotic recombination (Knudson, 1978). All of these mechanisms would convert a cell that is heterozygous for a mutation to a cell that is homozygous for loss of the normal allele. What might be the function of such a gene? The answer to this question would then identify the fundamental nature of the defect in a human tumor. Retinoblastoma is of particular value in this regard because it is the only tumor that exists in three forms, that is nonhereditary, hereditary without visible chromosomal change, and hereditary with visible chromosomal change. Here alone do we have an opportunity to make a chromosomal localization and pursue studies o f the function of such a gene.

Fortunately, there was new activity in pursuit of the gene. Robert Sparkes attempted to find a marker gene that could be used to compare tumor cells and normal cells from the same case, for evidence of a second event. He discovered to his surprise that just one other gene, esterase D, was known to be on chromosome 13. He found that deletion cases had only 50% of the normal amount of the enzyme in their blood cells, so that the two genes were near each other (Sparkes et al., 1980). He allso found a polymorphic, electrophoretic variant that was frequent enough to be useful in linkage analysis. In fact it was closely linked to the retinoblastoma gene, so that the deletion cases and nondeletion cases did involve the same gene (Sparkes et al., 1983). Then the groups of William Benedict and Sparkes found a case in which there was a 50% level of esterase D in the blood cells and none in the tumor (Benedict et ai!., 1983). This was direct evidence for the recessive hypothesis. At the same time, two Toronto investigators, Brenda Gallie and Robert Phillips, and their colleagues (Godbout et al., 1983) were investigating esterase D in tumors of patients whose blood cells were heterozygous for the two different electrophoretic alleles of esterase D; four of six tumors showed a loss of heterozygosity, with retention of just one allele. Here too was evidence for the recessive hypothesis. Meanwhile, Webster Cavenee and Raymond White had undertaken to find polymorphic DNA segments (RFLPs) throughout the human genome with the use of restriction enzymes. They undertook an examination of markers on chromosome 13. This approach had several advantages. There was a single assay method for all markers, whereas protein assays like that for esterase D were unique for each marker. In addition, it became possible to study multiple markers on one chromosome to gain knowledge about the mechanisms in a particular tumor. Thus, if a marker near the centromere retained heterozygosity and distal markers lost it but the amount of product remained the same, there was strong evidence that mitotic recombination had occurred. These investigators,

14

ALFRED G. KNUDSON, JR.

together with Gallie, Phillips, Benedict, Strong and colleagues, were able to find support for all of the proposed mechanisms of deletion, chromosomal loss, and recombination as second events (Cavenee et al., 1983). Furthermore, there was no difference between tumors associated with germ line mutation and those not so associated. Presumably the tumors that revealed no loss had sustained local mutations in or around the gene, but that could not be ascertained until the gene was cloned. Cavenee was subsequently able to demonstrate in a familial case that the retained polymorphic marker in one tumor was inherited from the affected parent; the marker from the unaffected parent had been lost, as expected (Cavenee et al., 1985). One investigator who expressed to me an early interest in the retinoblastoma problem was Sam Latt, who had developed a technique for the study of sister chromatid exchanges and was interested in RFLPs as a tool for the study of retinoblastoma. Working in his laboratory, Dryja isolated the RFLP, H3-8, that would lead to the cloning of the gene. Meanwhile, a young pediatrician, Stephen Friend, who even as a resident physician at the Children’s Hospital of Philadelphia expressed interest in retinoblastoma, took up a fellowship at Children’s Hospital in Boston. i n Weinberg’s laboratory he and Dryja attacked the problem with the use of H3-8, and the group cloned the RBI gene (Friend et al., 1986). Two other groups, one led by Wen-Hwa Lee (Lee et al., 1987a) and the other by Benedict (Fung et al., 1987), confirmed the discovery. Furthermore, Lee’s group found that the protein was phosphorylated and could interact with DNA (Lee et al., 1987b), and that the normal gene could reverse the transformation of retinoblastoma cells in uitro (Huang et al., 1988). T h e first human recessive cancer gene was in hand. By this time (1986), such genes were known by two names, antioncogenes and tumor suppressor genes. The former was a term I had introduced at a meeting in Gatlinburg, T N , in 1982 (Knudson, 1983), in an attempt to express the idea that the normal alleles were in effect antioncogenic, as well as the thought that they might be acting physiologically against the oncogenes. Other investigators preferred the term tumor suppressor genes, as a reflection of the idea that normal alleles could suppress tumorigenicity [reviewed by Klein ( 1987)]. B. RB1, TP53, A N D DNA TUMOR VIRUSES

By 1984 it was established that one of the DNA viruses, human papilloma \.irus (HPV), was the major causative agent for human cervical carcinoma. It thus joined the aforementioned DNA viruses as a subject of

MUTATION AND CANCER: A PERSONAL ODYSSEY

15

great interest. Oncogenic HPVs were like SV40 and oncogenic adenoviruses in that their DNAs were also integrated into the host’s genome. T h e oncogenic strains were found to have transforming genes, one, E6, being similar to E l B in its target and the other, E7, similar to ElA. E6 and E7 together were capable of transforming cells in vitro (Hawley-Nelson et al., 1989; Miinger et al., 1989). These oncogenes, like those of the other DNA tumor viruses, did not have any cellular homologues. By then the idea had developed that the viral oncogenic proteins were in some way interacting with host proteins to interfere with the regulation of DNA synthesis. Indeed, clear evidence was found for the formation of protein complexes between these viral proteins and host proteins of various molecular weights. One of the latter, p53 (Lane and Crawford, 1979; Linzer and Levine, 1979), complexed with T, ElB, and E6 and was mutant in certain tumors (DeLeo et al., 1979; Baker et al., 1989). T h e gene, tumor protein 53 (TPH), was in fact a tumor suppressor gene. It was also discovered that another binding protein, p105, was identical to the RBI protein product, pRB. It interacted with adenovirus E1A protein (Whyte et al., 1988), SV40 T protein (DeCaprio et al., 1988) and HPV E7 protein (Dyson et al., 1989). These three viruses had transformirfg genes whose products combined with, and apparently inactivated, two important antioncogenes. The latter were not mutant under these circumstances. Finally, the mechanism of transformation by DNA tumor viruses was illuminated. The RNA tumor viruses yielded oncogenes and pointed the way to cellular protooncogenes; the DNA viruses contained transformation genes whose products interacted with antioncogenes. I n each case, the DNA tumor virus inactivates a protein that is widely expressed and plays a critical role in regulating the cell cycle. T h e RBI protein in its unphosphorylated form interferes with passage through G1 to S in the cell cycle, and in its phosphorylated form it is permissive for such passage (Buchkovich et al., 1989; Chen et al., 1989; DeCaprio et al., 1989). T h e viral proteins interact with the unphosphorylated form. Passage through the S phase therefore can follow phosphorylation of pRB, mutation o r loss of RBI, or interaction of pRB with certain viral proteins. TP53 also proved to play a role in regulating the cell cycle, as Diller et al. (1990) suggested for osteosarcoma cells that were mutant for the gene. Production of p53 increases in cells in response to ionizing radiation, and cells are arrested in the G1 phase of the cell cycle (Kastan et al., 1991). Cells deficient in p53 do not arrest, proceed through the S phase, and incur chromosomal alterations. Evidently p53 can provide a signal for DNA repair before the next round of DNA synthesis; lacking such a

16

ALFRED G. KNUDSON, JR.

signal, cells proceed with damaged DNA. One kind of aberration is gene amplification, which has been employed to assess the effect of p53. Cells deficient in p53 show several orders of magnitude more amplification than d o those with normal p53 activity (Livingstone et al., 1992; Yin et al., 1992). Some kinds of cells apparently can remain in G1 for prolonged periods in response to DNA damage, if wild-type p53 is present; other cells may undergo apoptosis (Shaw et al., 1992; Yonish-Rouach et al., 1991; Lowe et al., 1993a,b; Clarke et al., 1993). TP53 thus seems to act in a critically conditional fashion, responding to certain signals of genomic aberration. In response to an increase in p53, a cell may be arrested in the G1 phase or directed to apoptosis. A study suggests a third alternative, i.e., switching from an exponential growth mode to a renewal growth mode (Sherley et al., 1995). This last response would be particularly applicable to epithelial tissues. C. GENETICEVENTSI N CARCINOGENESIS

From the genetic point of view, human cancers readily fall into two categories, according to the presence o r absence of specific initiating translocations. T h e former category embraces much of leukemia, lymphoma, and sarcoma, but seems to play no role in the common carcinomas. Evidently these translocations impart such a great growth advantage that for many tumors no other events seem to be necessary. However, other events can intervene, a classical example being chronic myelocytic leukemia (CML). In CML the chronic phase is associated with just the characteristic 9;22 translocation. In the acute blastic crisis, which typically occurs after 3-5 years, other genetic events appear. One of these is a second Ph'. Here a relatively benign and stable, although very abnormal, disorder gives way to a very malignant one. If the malignant phase of such a disease were to occur much earlier, the observer would find a cancer with more than one genetic lesion, as may be the case for some acute leukemias with Ph'. If the malignant phase were never to occur before death from other causes, the observer could find a nonprogressive condition with a single genetic lesion; this seems to happen in some cases of chronic lymphocytic leukemia with a translocation. For most cancers the initiating genetic lesion seems to be mutation o r loss of both copies of an antioncogene. T h e prototypic tumor, retinoblastoma, appears to have just one rate-limiting event, mutation or loss of both copies of RBI. However, other events are to be found in tumor cells. In fact, in one study 8 of 11 tumors revealed an iso(6p) chromosome (Kusnetsova et al., 1982). Such a change could be critical

MUTATION AND CANCER: A PERSONAL ODYSSEY

17

for tumor progression and metastasis, but recognition of the tumor may not depend upon this abnormality; only the two “hits” that involve RBI appear to be rate-limiting. T h e list of antioncogenes that are abnormal in hereditary cancers has grown quickly: a Wilms’ tumor gene, WTI (1990); neurofibromatosis type 1, NFI (1990); adenomatous polyposis coli, APC (1991); NF2 (1993); von Hippel-Lindau syndrome, VHL (1993); tuberous sclerosis 2 (1994); and hereditary breast cancer, BRCAI (1994). Furthermore, TP53 has been found to be mutant in the germ line in Li-Fraumeni syndrome (Malkin et al., 1990). In each case, tumors show loss or mutation of the second copy of the gene, as expected, and for every gene except BRCAI, at least some nonhereditary tumors of the same kind show similar aberrations. For some tumors, for example, some cases of Wilms’ tumor and the benign adenomatous polyp, one antioncogene appears to be ratelimiting, as for retinoblastoma. This is also true for very small renal tumors that appear in a rat model that we have been studying. This model of hereditary renal carcinoma in the rat (Eker, 1954), can now be attributed to mutation in an antioncogene, the rat homologue of the human tuberous sclerosis 2 gene (Yeung et al., 1994; Kobayashi et al., 1995). Only a few of the small tumors become large ones, and in the latter other genetic changes, on chromosome 5, are compatible with the loss of a second antioncogene, near the a-interferon complex (Hino et al., 1993). In the best defined sequence among human carcinomas, colon cancer, Bert Vogelstein’s group has characterized a succession of changes affecting APC, RAS, TP53, and the Deleted in Colon Cancer gene (DCC) (Fearon and Vogelstein, 1990). Mutation of APC in the germ line gives rise to familial adenomatous polyposis, which is strongly predisposing to colon cancer. It leads to a myriad of benign precursors, adenomatous polyps, which arise when the second copy of the gene is lost or mutant. T h e mutation in effect produces a great increase in the number of dividing target cells that can be affected by still further mutations. At the cellular level, transformation of a cell in an adenomatous polyp into a carcinoma cell is an extremely rare event. It now seems probable that loss or mutation of both TP53 genes transforms cells (Kikuchi-Yamoshita et al., 1992). Once both normal TP53 genes are lost, other genetic aberrations can occur quickly, so that they do not appear to be rate-limiting. Carcinoma arises in two clinically apparent stages, the benign polyp and the cancer. It seems that each of these can be considered to result from two hits, acknowledging that the other events that follow, but that are not rate-limiting, may be pathologically important.

la

ALFRED G. KNUDSON, JR.

Germ line mutation of TP53 predisposes to breast carcinoma, but not notably to colon carcinoma or small cell carcinoma of the lung. Yet, somatic mutations of TP53 are more common in the latter two tumors than in the former. It is interesting that several of the tissues that are the strongest targets in the Li-Fraumeni syndrome are those that undergo growth in adolescence. Presumably this growth involves an increase in the number of stem cells. It seems that other events, such as mutation of APC, accomplish this in other epithelial tissues. This requirement for other events diminishes the impact of the inherited mutation (Knudson, 1992). A similar phenomenon is observed for R B I . This gene is commonly mutated in certain sarcomas and carcinomas, especially small cell carcinoma of the lung. As with TP53 in the Li-Fraumeni syndrome, germ line mutation of RBI in hereditary retinoblastoma predisposes to sarcomas but not to small cell carcinoma of the lung.

IV. Conclusions During the professional lifetimes of many living scientists, growing knowledge of biological science, notably genetics, virology, and molecular biology, has taken us from a state of ignorance of the composition of the gene itself to a state of considerable understanding of the nature and number of critical genetic events in human carcinogenesis. At the time, some 50 years ago, that DNA was being shown to be the genetic material, two ideas on the origin of cancer were prevalent. Since then, discoveries on tumor viruses and somatic mutations have combined to produce a new view of carcinogenesis in humans. T h e RNA tumor viruses revealed oncogenes and, in turn, their host counterparts, protooncogenes. Cytogenetic analysis led to the discovery of specific somatic mutations in these same protooncogenes, apparently unrelated to viral infection. T h e DNA tumor viruses revealed tumor antigen genes that had no counterparts in their hosts. The investigation of hereditary cancers, and then of their corresponding nonhereditary forms, led to the discovery of antioncogenes, or tumor suppressor genes; two of these interact with the tumor antigens of DNA tumor viruses. Returning to the initial theme of this review, it is of considerable interest that two tumors of children, Burkitt’s lymphoma and retinoblastoma, have played key roles in the development of our understanding of genetic changes in human cancer. Since then, notions of the origin of cancer that seemed so diverse have fused into a singular view. Even though much is yet to be learned, there is a strong sense that cancer can be explicated and that new measures for prevention and treatment can be generated with that understanding.

MUTATION AND CANCER: A PERSONAL ODYSSEY

19

ACKNOWLEDGMENTS The author acknowledges financial support by a grant from the Lucille P. Markey C:haritable Trust, a grant (CA-06927) from the National Cancer Institute, and an appropriation from the Commonwealth of Pennsylvania.

REFERENCES

Ames, B. N., Durston, W. E., Yamasaki, E., and Lee, F. D. (1973). Proc. Natl. Acad. Sci. USA 70, 2281-2285. Armitage, P., and Doll, R. (1954). Br.J. Cancer 8, 1-12. Armitage, P., and Doll, R. (1957). Br.J. Cancer 11, 161-169. Bader, J. P. (1964). Virology 22, 462-468. Baker, S. J., Fearon, E. R., Nigro, J. M., Hamilton, S. R., Preisinger, A. C., Jessup, J. M., van Tuinen, P., Ledbetter, D. H., Barker, D. F., Nakamura, Y., White, R., and Vogelstein, B. (1989). Science 244, 21 7-22 1 . Baltimore, D. (1970). Nature (London) 226, 1209-121 1. Elaluda, M. A,, and Goetz, I. (1961). Virology 15, 185-199. Ebaluda, M. A., and Jamieson, P. P. (1961). Virology 14, 33-45. Elaluda, M. A., and Nayak, D. P. (1970). Proc. Natl. Acad. Sci. USA 66, 329-336. Benedict, W. F., Murphree, A. L., Banerjee, A,, Spina, C. A., Sparkes, M. C., and Sparkes, R. S. (1983). Science 219, 973-975. Berenblum, I. (1941). Cancer Res. 1, 807-814. Berenblum, I., and Shubik, P. (1947). Br.1. Cancer 1, 383-391. Black, P. H., Rowe, W. P., Turner, H. C., and Huebner, R. J. (1963). Proc. Natl. Acad. Sci. USA 50, 1148-1 156. Boveri, T. (1914). “Zur Frage der Entstehung Maligner Tumoren,” pp. 1-64. Fischer, Jena. Eirookes, P., and Lawley, P. D. (1964). Nature (London) 202, 781-784. Elrugge, J. S., and Butel, J. S. (1975).J. Virol. 15, 619-635. Elurger, H., and Doerfler, W. (1974).J. Virol. 13, 975-992. ELuchkovich, K., Duffy, L. A., and Harlow, E. (1989). Cell 58, 1097-1 105. Cavenee, W. K., Dryja, T. P., Phillips, R. A., Benedict, W. F., Godbout, R., Gallie, B. L., Murphree, A. L., Strong, L. C., and White, R. L. (1983). Nature 305, 779-784. Cavenee, W. K., Hansen, M. F., Nordenskjold, M., Kock, E., Maumenee, I., Squire, J. A., Phillips, R. A., and Gallie, B. L. (1985). Sczence 228, 501-503. Chen, P.-L., Scully, P., Shew, J.-Y., Wang, J. Y. J., and Leed, W.-H. (1989). Cell 58, 11931198. Clarke, A. R., Purdie, C. A., Harrison, D. J., Morris, R. G., Bird, C. C., Hooper, M. L., and Wylie, A. H. (1993). Nature (London) 362, 849-852. Cleaver, J. E. (1968). Nature (London) 218, 652-656. Collins, S., and Groudine, M. (1982). Nature (London) 298, 679-681. Comings, D. (1973). Proc. Natl. Acad. Sci. USA 70, 3324-3328. 1)alla-Favera, R., Bregni, M., Erikson, J., Patterson, D., Gallo, R. C., and Croce, C. M. (1982a). Proc. Natl. Acad. Sci. USA 79, 7824-7827. Dalla-Favera, R., Wong-Staal, F., and Gallo, R. C. (1982b). Nature (London) 299, 61-63. Dargeon, H. W. (1960). “Tumors of Childhood.” Hoeber, New York. DeCaprio, J. A., Ludlow,J. W., Figge, J., Shew, J.-Y., Huang, C.-M., Lee, W.-H., Marsilio, E., Paucha, E., and Livingston, D. M. (1988). Cell 54, 275-283.

20

ALFRED G . KNUDSON, JR.

DeCaprio, J . A.. Ludlow, J. W., Lynch, D., Furukawa, Y. O., Griffin, J., Piunica-Worms, H., Huang, C.-M., and Livingston, D. M. (1989). Cell 58, 1085-1095. DeKlein, A,, Van Kessel, A. G., Grosveld, G., Bartram, C. R., Hagemeijer, A., Bootsma, D., Spurr, N. K., Heisterkarnp, N., Groffen, J., a n d Stephenson, J. R. (1982). Nature (London) 300, 765-767. DeLeo, A. B., Jav. G . , Appella, E., DuBois, G . C., Law, L. W., and Old, L. J. (1979). Proc. Natl. Acud. Sct. LISA 76, 2420-2424. Der, C. J., Krontiris, T. G., and Cooper, G. M. (1982). Proc. Natl. Acud. Sci. USA 79, 36373640. Diller, L., Kassel, J., Nelson, C. E., Gryka, M. A,, Litwak, G., Gebhardt, M., Bressac, B., Ozturk, M.. Baker, S. J., Vogelstein, B., and Friend, S. H . (1990). Mol. Cell Bzol. 10, 5772-5781. Duesberg, P. H . , a n d Vogt, P. K. (1970). Proc. Natl. Acud. Sci. USA 67, 1673-1680. Dyson, N., Howley, P. M., Munger, K., and Harlow, E. (1989). Science 243, 934-937. Eker, R. (1954). act^ Pathol. Microbiol. Sca7d. 34, 554-562. Erikson, J., ar-Rushdi, A,, Drwinga, H . L., Nowell, P. C., and Croce, C. M. (1983). Proc. Natt. A d . Set. USA 80, 820-824. Falls, H . F., a n d Neel, J. V. (1951). Arch. Ophthalmol. 46, 367-389. Fearon, E. R., and Vogelstein, B. (1990). Cell 61, 759-767. Fialkow, P. J., Gartler, S. M., and Yoshida (1967). Proc. Natl. Acad. Sci. USA 58, 1468-1471. Foulds, L. (1954). Cancer Res. 14, 327-339. Francke, U., and Kung, F. (1976). Med. Pediafr. Oncol. 2, 379-385. Fried, M. (1965). Proc. 'Vatl. Acud. Sci. USA 53, 486-491. Friedewald, W. F., and Rous, P. (1944).J. Exp. Med. 80, 101-125. Friend, S. H., Bernards, R., Rogelj, S., Weinberg, R. A., R a p p o r t ,J. M., Albert, D. M., and Dryja, T. P. (1986). Nature 323, 643-646. Fung, Y.-K. T., Murphree, A. L., T'Ang, A., Qian, J., Hinrichs, S. H., and Benedict, W. F. (1987). Science 236, 1657-1661. Godbout. R., Dryja, T. P., Squire, J., Gallie, B. L., and Phillips, R. A. (1983). Nature 304, 451-453. Graham, F. L., Abrahams, P. J.. Mulder, C., Heiljneker, H. L., Warnaar, S. O., d e Vries, F. A. J., Fiers, W.. and van d e r Eb, A. J. (1975). Cold Spring Harbor Symp. @ant. Biol. 39, 637-650. Habel, K. (1962). Cold Spring Harbor Symp. Quant. Biol. 27, 433-439. Hamlyn, P. H.. and Rabbitts, T. H . (1983). Nuture (London) 304, 135-139. Harris, H., Miller, 0.J., Klein, G . ,Worst, P., and Tachibana, T. (1969).Nature(Lond0n) 223, 363-368. Hawley-Nelson, P., Vousden, K. H., Hubbert, N. L., Lowy, D. R., and Schiller, J. T. (1989). E M B O J . 8, 3905-3910. Hayward, W. S . , Neel, B. G., and Astrin, S. M. (1981). Nature (London) 290,475-480. Hethcote, H. W., and Knudson, A. G. (1978). Proc. Natl. Acad. Sci. USA 75, 2453-2457. Hino, O., Klein-Szanro, A. J. P., Freed, J. J., Testa, J. R., Brown, D. Q.,Vilensky, M., Yeung, R. S., Tartof, K. D., and Knudson, A. G. (1993).Proc. Natl. Acad. Scz. USA 90,327-331. Hirai, K., and Defendi, V. (1972).J . Vzrol. 9, 705-707. Huang, H-.J. S., Yee, 1.-K., Shew, J.-Y., Chen, P.-L., Bookstein, R., Freidman, T., Lee, E. Y.-H. P., and Lee, W.-H. (1988). Science 242, 1563-1566. Huebner, R. J., and Todaro, G. J. (1969). Proc. "Wl. Acad. Scz. U S A 64,1087-1094. Huebner, R. J., Rowe, W. P., Turner, H . C., and Lane, W. T. (1963). Proc. Natl. Acad. Sci. USA 50, 379-389.

MUTATION AND CANCER: A PERSONAL ODYSSEY

21

Kastan, M. B., Onyekwere, O., Sidransky, D., Vogelstein, B., and Craig, R. W. (1991). Cancer Res. 51, 6304-631 1. K.ikuchi-Yamoshita, R., Konishi, M., Ito, S., Seki, M., Tanaka, K., Maeda, Y., Iino, H., Fukayama, M., Koike, M., Mori, T., Sakuraba, H., Fukunari, H., Iwama, T., and Miyaki, M. (1992). Cancer Res. 52, 3965-3971. Hein, G. (1981). Nature (London) 294, 313-318. Klein, G. (1987). Science 238, 1539-1544. Knudson, A. G. (1965a). Cancer 18,815-818. Knudson, A. G. (1965b). “Genetics and Disease.” McGraw-Hill, New York. Ehudson, A. G. (1966). Am. Naturalist 100, 162-164. Knudson, A. G. (1971). Proc. Natl. Acad. Sci. USA 68, 820-823. Knudson, A. G. (1973). Adu. Cancer Res. 17, 317-352. Knudson, A. G. (1978). Semin. Oncol. 5, 57-60. Knudson, A. G. (1983). Prog. Nucleic Acid Res. Mol. Biol. 29, 17-25. Knudson, A. G. (1992). Sem. Cancer Biol. 3,99-106. Knudson, A. G., and Arnromin, G. D. (1966). Cancer 19, 1032-1037. Knudson, A. G., and Strong, L. C. (1972a). Am. J. Human Genet. 24,514-532. Knudson, A. G., and Strong, L. C. (1972b).J. Natl. Cancer Inst. 48, 313-324. Knudson, A. G., Brodetsky, A. M., and Baluda, M. A. (1967).J. Virol. 1, 1150-1157. Knudson, A. G., Hethcote, H. W., and Brown, B. W. (1975). Proc. Natl. Acad. Sci. USA 72, 5 116-5 120. Knudson, A. G., Meadows, A. T., Nichols, W. W., and Hill, R. (1976).New Engl. J. Med. 295, 1120-1 123. Kobayashi, T., Hirayama, Y., Kobayashi, E., Kubo, Y., and Hino, 0. (1995).Nature Genet. 9, 70-74. Koprowski, H., Ponten, J. A., Jensen, F., Ravdin, R. G., Moorhead, P., and Saksela, E. (1962).J. Cell. Comp. Physiol. 59, 281-292. Kusnetsova, L. E., Prigogina, E. L., Pogosianz, H. E., and Belkina, B. M. (1982). Human Genet. 61, 201-204. Lacour, F., Delain, E., Gerard-Marchant, R., and Weiler, 0. (1970). C. R. Acad. Sci. (0) (Paris) 271, 141-144. Ladda, R., Atkins, L., Littlefield,J., Neurath, P., and Marimuthu, K. M. (1974).Science 185, 784-787. Land, H., Parada, L. F., and Weinberg, R. A. (1983). Nature 304, 596-602. Lane, D. P., and Crawford, L. V. (1979). Nature (London) 278, 261-263. Lee, W.-H., Bookstein, R., Hong, F., Young, L.-J., Shew,J.-Y., and Lee, E. Y.-H. P. (1987a). Science 235, 1394- 1399. Lee, W.-H., Shew, J.-Y., Hong, F. D., Sery, T. W., Donoso, L. A., Young, L.-J., Bookstein, R., and Lee, E. Y.-H. P. (1987b). Nature 329, 642-645. Lele, K. P., Penrose, L. S., and Stallard, H. B. (1963). Ann. Human Genet. 27, 171-174. Linzer, D. I., and Levine, A. J. (1979). Cell 17, 43-52. Livingstone, L. R., White, A., Sprouse, J., Livanos, E., Jacks, T., and Tlsty, T. D. (1992). Cell 70,923-935. Lowe, S. W., Ruley, H. E., Jacks, T., and Housman, D. E. (1993a). Cell 74, 957-967. Lowe, S. W., Schmitt, E. M., Smith, S. W., Osborne, B. A., and Jacks, T. (1993b). Nature (London) 362, 847-849. Malkin, D., Li, F. P., Strong, L. C., Fraumeni, J. F., Nelson, C. E., Kim, D. H., Kassel, J., Gryka, M. A., Bischoff, F. Z., Tainsky, M. A., and Friend, S. H. (1990). Science 250, 1233-1238,

22

ALFRED G. KNUDSON, JR.

Manolov, G., and Manolova, Y. (1972). Nature (London) 237, 33-34. Manolova, Y., Manolov, G., Kieler, J., Levan, A , , and Klein, G. (1979). Hereditas 90, 5-10. Martin, G. S. (1970). Nature (London) 227, 1021-1023. Martin, R. G., and Chou, J. Y. (1975).J. Virol. 15, 599-612. Miller, E. C., Miller, J. A,, and Hartmann, H. A. (1961). CancerRes. 21, 815-824. Miller, R. W., Fraumeni, J. F., and Manning, M. D. (1964). New Eng1.J. Med. 270,922-927. Moolgavkar, S. H., and Knudson, A. G. (198I).J. Nad. Cancer Inst. 66, 1037-1052. Muller, H. J. (1927). Sciencr 66, 84-87. Munger, K., Phelps, W. C., Bubb, V., Howley, P. M., and Schlegel, R. (1989).J. Virol. 63, 4417-442 1. Nordling, C. 0. (1953). Er. J . Cancer 7 , 68-72. Nowell, P. C. (1993).A h . Cancer Res. 62, 1-17. Nowell, P. C., and Hungerford, D. A. (1960). Science 132, 1497. Osborn, M., and Weber, K. (1975).J. viral. 15, 636-644. Parada, L. F., Tabin, C. J., Shih, C., and Weinberg, R. A. (1982). Nature (London) 297,474478. Ponnamperuma, C. A,, Lemmon, R. M., and Calvin, M. (1962). Science 137, 605-607. Pope, J. H., and Rowe, W. P. (1964).J. Exp. iMed. 120, 577-588. Rabbitts, T. H. (1994). Nature (London) 372, 143-149. Riccardi, V. M.,Sujansky, E., Smith, A. C., and Francke, U. (1978). Pdiatrics 61,604-610. Rosenthal, P. N., Robinson, H. L., Robinson, W. S., Hanafusa, T., and Hanafusa, H. (1971). Proc. Natl. Arad. Sci. USA 68, 2336-2340. Rous, P., and Kidd, J. G. (1941).J. Exp. Med. 73, 365-390. Rowley, J. D. (1973). Nature (London) 243, 290-293. Ruley, H. E. (1983).Nature (London) 804, 602-606. Sambrook, J., Westphal, H., Srinavasan. P. R., and Dulbecco, R. (1968). Proc. Natl. Acad. Sci. USA 60, 1288- 1295, Santos, R., Tronick, S. R., Aaronson. S. A., Pulciano, S., and Barbacid, M. (1982). Nature (London) 298,343-347. Schappert-Kimmijser, J., Hemmes, G. D., and Nijland, R. (1966). Ophthalmologica 151, 197-2 13. Schwab, M., Alitalo, K., \’armus, H. E., Bishop, J. M., and George, D. (1983). Nature (London) 303,497-501. Sharp, P. A., Gallimore, P. H., and Flint, S. J. (1975). Cold Spring Harbor Symp. Quant. Eiol. 39,457-474. Shaw, P., Bovey, R., Tardy, S., Sahli, R., Sordat, B., and Costa, J. (1992). Proc. Natl. Acad. Sci. USA 89,449.5-4499. Shein, H. M., and Enders, J. F. (1962). Proc. h’atl.A d . Sci. USA 48, 1 164-1 172. Sherley, 3. L,., Stadler, P. B., and Johnson, D. R. (1995). Proc. Natl. Acad. Sri. C S A 92, 136140. Shih, C . , Shilo, B.-Z., Goldfarb, M. P., Dannenberg, A,, and Weinberg, R. A. (1979). Proc. Natl. A d . Sci. USA 76, 5714-5718. Sparkes, R. S., Sparkes, M. C., Wilson, M. G . , Towner, J. W., Benedict, W., Murphee, A. L., and Yunis, J. S. (1980). Science 208, 1042-1044. Sparkes, R. S.,Murphee, A. L., Lingua, R. W., Sparkes, M. C., Field, L. L., Funderburk, S. J . , and Benedict, W. F. (1983). Science 219, 971-973. Stanbridge, E. J. (1976). Nature (London) 260, 17-20. Stehelin, D., Varmus, H. E., Bishop. J. M., and Vogt, P. K. (1976). Nature (London) 260, 170-1 73.

MUTATION AND CANCER: A PERSONAL ODYSSEY

23

Summers, J., and Mason, W. S. (1982). Cell 29, 403- 415. l h b , R., Kirsch, I., Morton, C., Lenoir, G., Swan, D., Tronick, S., Aaronson, S., and Leder, R. (1982). Proc. Natl. Acad. Sci. USA 79, 7837-7841. Trgtmeyer, P. (1975).J. Virol. 15, 613-618. Temin, H. M. (1963). Virology 20, 577-582. Temin, H. M. (1964a). Virology 23, 486-494. Temin, H. M. (1964b). Proc. Natl. Acad. Sci. USA 52, 323-329. Temin, H. M., and Mizutani, S. (1970). Nature ( L o d o n ) 226, 121 1-1213. Toyoshima, K., and Vogt, P. K. (1969). Virology 39, 930-931. Vogt, M., and Dulbecco, R. (1960). Proc. N d l . Acad. Sci. USA 46, 365-370. Vogt, M., and Dulbecco, R. (1963). Proc. Natl. Acad. Sci. USA 49, 171-179. \Tang, L. H., Duesberg, P., Beemon, K., and Vogt, P. K. (1975).J. Virol. 16, 1051-1070. Whyte, P., Buchkovich, K. J., Horowitz,J. M., Friend, S. H., Raybuck, M., Weinberg, R. A., and Harlow, E. ( 1988). Nature 334, 124- 129. Yleung, R. S., Xiao, G.-H.,Jin, F., Lee, W.-C., Testa, J. R., and Knudson, A. G. (1994). Proc. Natl. Acad. Sci. USA 91, 11413-11416. Yin, Y., Tainsky, M. A,, Bischoff, F. Z., Strong, L. C., and Wahl, G. M. (1992). Cell 70,937948. Yonish-Rouach, E., Resnitsky, D., Lotem, J., Sachs, L., Kimchi, A., and Oren, M. (1991). Nature (London) 352, 345-347. Zech, L., Haglund, V., Nilsson, K., and Klein, G . (1976). 1nt.J. Cancer 17, 47-56.

This Page Intentionally Left Blank

ONCOGENIC ROLE OF “MASTER” TRANSCRIPTION FACTORS IN HUMAN LEUKEMIAS AND SARCOMAS: A DEVELOPMENTAL MODEL A. Thomas Look Department of Experimental Oncology, St. Jude Children’s Research Hospital, and University of Tennessee College of Medicine, Memphis, Tennessee 38105

I. Introduction 11. Transcriptional Control Genes Altered by Chromosomal Rearrangements A. Acute Lymphoid Leukemias B. Acute Myeloid Leukemias C. Sarcomas ILII. Oncogenic Transcription Factors and the Developmental Regulatory Proteins of Drosophila A. Homologies within DNA-Binding and Dimerization Domains B. Roles of Oncogenic Transcription Factors in Normal Mammalian Development C. Oncogenic Transcription Factors May Disrupt Normal Regulatory Networks That Determine Hematopoietic and Mesenchymal Cell Fate IV. Summary and Future Directions References

1. Introduction

Abnormally expressed proteins can induce cancer by mimicking the action of growth factors or their receptors, signal transducers, or transcription factors, which affect gene expression directly. Oncogenesis mediated by transcription factors has particular significance in the acute leukemias and sarcomas, where chromosomal translocation and inversions commonly activate genes encoding nuclear DNA-binding proteins (Rowley et al., 1993; Rabbitts, 1994). Recent studies of the structural features and binding patterns of these gene products have afforded pivotal insights into the mechanisms by which transcription factors could participate in tumorigenesis. These versatile proteins bind to regulatory elements (promoters and enhancers) in DNA, through which they stimulate or inhibit gene transcription (Papavassiliou, 1995). The modular organization of transcription factors into distinct domains provides an ideal system for their multiple functions: binding to DNA, transactiva25 ADVANCES I N C A N C E R RESEARCH. VOL. 67 All

Copyright 0 1995 hy Academic Press, Inc. of reproduction i n any form reserved.

righi5

26

A. THOMAS LOOK

tion of target genes, and protein-protein interactions within complex regulatory networks. Reassortment of these domains by chromosomal rearrangement has the potential to render the resulting hybrid transcription factors tumorigenic. Most intriguing perhaps are the similarities between conserved regions of mammalian transcription factors ’ and those of “master” developmental proteins regulating the earliest stages of Drosophila embryogenesis (Nusslein-Volhard and Wieschaus, 1980; Nusslein-Volhard et al., 1987; Levine and Harding, 1989). In this chapter, I review recent findings in this emerging area of research and attempt to integrate them into a developmental model of progenitor cell oncogenesis on the basis of disruption of transcriptional control.

I I . Transcriptional Control Genes Altered by Chromosomal Rearrangements

Transcription factor genes are often disrupted by chromosomal translocations in human acute leukemias and sarcomas; the resulting altered o r dysregulated proteins, which retain the ability to bind DNA and interact with other factors, contribute to the differentiation arrest and aberrant growth of specific hematopoietic and mesenchymal progenitor cells. These nuclear proteins possess highly conserved amino acid sequence motifs within their DKA- and protein-binding domains, permitting their classification by common structural features, exemplified by basic helix-loop-helix (bHLH), basic region-leucine zipper (bZIP), cysteine-rich LIM, homeodomain (HOX), zinc finger, A-T hook, and Ets-like and runt homology domains (Tables I and 11). Transcription factor genes appear to be the preferred targets of gene rearrangements in particular morphological types of hematopoietic and mesenchymal cells, suggesting that their transforming potential (hence, specificity) differs according to the stage of cell development. A unifying hypothesis that would account for the tumorigenicity of these proteins has been advanced by T. H . Rabbitts, who emphasizes similarities between oncogenic transcription factors and the products of “master genes” that specify lineage-specific patterns of gene expression during development (Rabbitts, 1991). By acting positively to up-regulate critical target genes or negatively to interfere with normal regulatory pathways, such factors can disrupt gene regulatory cascades that coordinate the expression of large numbers of proteins required for the completion of specific cell differentiation programs. Chromosomal translocations can alter transcription factors by more than one mechanism. For instance, the promoter-enhancer sequences of a gene on one chromosome can be moved near a gene encoding a

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION

FACTORS

27

potentially oncogenic transcription factor on a separate chromosome. T h e adjacent and rearranged promoter-enhancer elements then cause aberrant expression of the regulatory factor, leading to interference with the differentiation of hematopoietic or mesenchymal progenitors within a given cell lineage. A second (and perhaps the major) action of chromosomal translocations is gene fusion, whereby the DNA-binding, dimerization, and trans-effector regions of discrete genes are “stitched together” to produce a chimeric transcription factor with altered function. Although they share some features, the transcription factors invlolved in the acute lymphoid and myeloid leukemias and in the sarcomas possess unique properties that appear to restrict their modes of tumor initiation, as described in the following. A. ACUTELYMPHOID LEUKEMIAS

T h e immunoglobulin (Ig) and T-cell receptor (TCR) genes of normal lymphoid cells undergo diverse, clonal rearrangements, followed by highly regulated proliferation of the cells that successfully completes tlhese genetic changes. Hence, the subsequently generated B cells and T cells possess the specificities needed to support a competent immune system. Acute lymphoid leukemia (ALL) is often the product of translocations that affect the Ig, TCR, or other genes of lymphoid progenitors, eventually leading to dysregulated proliferation and clonal expansion of the transformed blast cells. The pathogenesis of ALL tends to preserve the expression of proteins that help to generate and maintain the normal phenotypes of developing lymphoid cells, although in socalled “mixed-lineage” cases there is aberrant expression of genes representing two or more lineages. Despite this evidence of disordered patterns of gene expression, ALL cells retain most of the features of normal lymphoid progenitors, and these immunophenotypic markers of progenitor cell lineage correlate remarkably well with specific types of chromosomal translocations and their unique reassortments of transcription factor genes (Table I and Fig. 1). In B-cell acute leukemia (also Burkitt’s lymphoma), the predominant t(8;14)(q24;q32) translocation rearranges one allele of MYC, a protlotypic basic region helix-loop-helix-zipper (bHLHzip) transcription factor gene on chromosome 8, into the heavy-chain Ig gene on chromosome 14q32, adjacent to the coding sequences of the Ig constant region (Dalla-Favera et al., 1982; Taub et al., 1982; Adams et al., 1983). This event results in dysregulation of MYC expression through the introduction of a strong Ig enhancer; mutations that accompany the t(8;14) are also important [reviewed in Rabbitts and Boehm (1991)l. In

Aflec ted grne MY C

Disease

References

Burkitt's lymphonia and B-cell ALL

Dalla-Favera et al. ( 1 982); Taub pt al. (1982); Adams rt al. (1983) Emanuel el al. (1984); Erikson et al. (1983); Rappold et al. (1984): Taub el ( ~ 1 .( 1 984) Hollis rt al. (3984); Croce et al. ( 1983) Finger et a1. ( I 986); McKeithan el al. (1986); Shima et a1. (1986)

M YC

I3asic. hrlix-loop-helix (bHLH) pt-oteins

t(X;22)(qL'4;qI1 )

MYC

t(8; I4)(q24;q I I )

M YC

T-cell ALL

l ( 7 ; IY)(q35;p13)

LYLI TALI ISCL I TCL5

T-cell ALL T-cell ALL

TAL2 RBTNI IT(;I

T-cell ALL.

t(ll;l4)(pl3;qll) t(7; I l)(q35;p13)

RB TN21TTG2 RBTN2ITTG2

T-cell ALL T-cell ALL

t( 10;14)(q24;qlI)

t(7;10)(q%;q24)

HOXl 1 HOXI I

T-cell ALL T-cell ALL

t(1; 19)(q23;p13)

E2A-PBX 1

Pre-B-cell ALL

t( 1;

14)(p32;qI 1)

t(7;9)(q35434) Cystein-rich ( L A M ) proteins

Honieodomain (HOX) proteins

t(1

1;14)(p15;qI1)

T-cell ALL

Mellentin et al. (1989) Begley et al. (1989); Chen rt al. ( 1990) Xia el al. (1991) McGuire et al. (1989); Greenberg et al. (1990) Boehm et al. (1991) Royer-Pokora et al. (1991) Hatano et al. (1991) Kennedy et al. (1991); Lu et al. (1991); Dube et al. (1991) Kamps et al. (1990);Nourse el al. ( 1990)

Rasic regiGn-!eu&e zipper (bZIP) proteins

.,,-l.,n\,-oo.no\ JnyLL,P'Ji

Zinc finger proteins

t( 15; 17)(q21;ql1-22)

PML-RAR'

APML

t(11;17)(q23;q21) t(3;v)(q26;v)

PLZF-RA R C EVIl

t(4; 1l)(q2 l;q23) t(9;l l)(p21;q23) t(l1;19)(q23;p13)

MLL-PBMI ClHRXFEL'IALLI -AF4< MLLIHRXIALLI-AF9c MLLIHRXIALL-ENLc

APML AML Pro-B-cell ALL AML(monocytic) ALL o r AML

t(8;21)(q22;q22)

AMLl -ETO'

AML (granulocytic)

AMLI-EVIL AML-EAP' CBFp-MYHl I

CML, blast crisis Myelodysplasia AML (myelomonocytic)

FUS-ERG'

AML

A-T hook minor groove binding proteins

Runt homology domain proteins N

W

Ets homology domain proteins

inaba et ai. (1992); Hunger et al. ( 1992)

L\1), 1

t(16;21)(pll;q22)

de T h e et al. (1990, 1991); Borrow et al. (1990); Longo et al. (1990); Kakizuka et al. (1991) Chen et al. (1993) Morishita et al. (1990, 1992) Morrissey et al. (1993); Gu et al. (1992); Domer et al. (1993) Nakamura et al. (1993) Tkachuk et al. (1992) Miyoshi et al. (1991); Gao et al. (1991); Erickson et al. (1992); Shimizu et al. (1992) Mitani et al. (1994) Nucifora et al. (1993) Liu et al. (1993) Ichikawa et al. (1994)

Abbreviations: AML, acute myeloid leukemia; ALL, acute lymphoblastic leukemia; APML, acute promyelocytic leukemia; CML, chronic myelogenous leukemkia. Based on the DNA-binding domain or protein-protein interaction domain included in the dysregulated or chimeric protein found in leukemic cells. CBFp does not have a runt homology domain, but binds as a complex with AMLl. Fusion gene. n

b

30

A . THOMAS LOOK

Pro-T Pluripotent

Lymphoid

Promonocyte

LYLl. TAL1, TALP, RBTN1, RBTNP

Monocyte

MLL-AFT

FIG. 1. Cell lineage-specific involvement of oncogenic transcription factors in the acute leukemias. The master developmental genes disrupted by chromosomal rearrangements in acute leukemia have shown a striking correspondence between specific gene rearrangements and defined morphologic and inimunophenotypic types of transformed lymphoid and myeloid progenitor cells, indicating the action of these dysregulated or hybrid transcription factors in disrupting highly specialized programs of hematopoietic cell differentiation. A major exception to this principle is the iMLL gene on chromosome 11, band q23, which is altered in leukemias of pro-B-cell lymphoid as well as myeloid progenitors, suggesting a regulatory role in both the lymphoid and myeloid developmental pathways. Asterisks refer to chimeric fusion genes (see Table I).

variant translocations [t(2;8) and t(8;22)], the MYC gene remains on chromosome 8, and portions of the K or A light-chain genes on chromosome 2 o r 22 are translocated to a site downstream of the MYC locus, again leading to aberrant expression of that gene (Emanuel et al., 1984; Erikson et al., 1983; Hollis et al., 1984; Rappold et al., 1984; Croce et al., 1983; Taub et al., 1984). How does MYC transform cells? T h e prevailing view suggests a transcriptional network of proteins, each possessing a bHLH as well as a leucine zipper protein dimerization motif. Thus, MYC is able to heterodimerize with the MAX protein (Blackwood and Eisenman, 1991; Prendergast et al., 1991), which in turn binds to DNA, to itself (MAX:MAX homodimers), and to other proteins, including MAD and Mxi- 1 (Ayer et

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

31

al., 1993; Zervos et al., 1993). Since only MYC:MAX heterodimers are ti-anscriptionally active (Blackwood and Eisenman, 1991; Prendergast et al., 1991), and since MYC and MAD have equal affinities for the MAX protein (Ayer and Eisenman, 1993; Larsson et al., 1994), increased MYC expression due to chromosomal rearrangement is thought to disrupt the MAX heterodimer equilibrium in lymphoid progenitors, leading to inappropriate transcription of downstream targets and ultimately to malignant transformation (and, perhaps, to the high proliferative rate characteristic of these neoplasms) (Amati et al., 1993). Support for this hypothesis comes from the induction of B-cell neoplasms in transgenic mice that carry the MYC oncogene driven by an Ig gene enhancer (Adams et al., 1985; Langdon et al., 1986). An activated MYC gene also induces tumorigenic conversion when it is introduced into human Epstein-Barr virus-infected B lymphoblasts in vitro (Lombardi et al., 1987). T h e role of transcription factors as the preferred targets of chromosomal translocations extends to acute T-cell leukemias, where the chromosomal break points consistently appear near enhancers included in the TCR P-chain locus on chromosome 7, band q34, or the ah-locus on chromosome 14, band q l 1. Highly active in committed T-cell progenitors, these enhancers stimulate strategically translocated transcription factors that are involved in cell differentiation but are not normally expressed in T lymphoid cells (Table I). The bHLH class of transcription factor genes, including MYC (Finger et al., 1986; McKeithan et al., 1986; Shima et al., 1986),TALl ISCL (Begley et al., 1989; Chen et al., 1990; Xia t t al., 1991), and LYLl (Mellentin et al., 1989), contains notable examples, in that the bHLH domains of their encoded proteins mediate dirnerizat ion and sequence-specific binding to DNA in the promoter-enhancer regions of target genes. A useful model is provided by TALI gene activation due to t(1;14) or to intragenic deletion on the 5’ side of the gene, rearrangements that affect up to one-fourth of all cases of childhood ‘T-cell leukemias (Baer, 1993). Since the TALl protein can dimerize with lE4’7 to form DNA-binding complexes (Hsu et al., 1991), its ectopic expression in T cells bearing t( 1;14) would be expected to activate specific $jetsof target genes that are normally quiescent in T-lineage progenitors. Several other genes whose products are transcriptional regulatory proi.eins can be altered by rearrangement into the vicinity of T-cell receptor 110ciinclude those encoding the cysteine-rich (LIM) proteins RBTN 1 ,and RBTN2 (also called T T G l and TTG2) (McGuire et al., 1989; Greenlberg et al., 1990; Boehm et al., 1991; Royer-Pokora et al., 1991).Although ithey contain LIM domains, which include zinc finger-like structures (Perez-Alvarado et al., 1994), these proteins lack homeobox DNA-binding

32

A . THOMAS LOOK

domains, a feature comnion to other transcription factors in this family. This suggests that the LIM domain may function in protein-protein rather than protein-DNA interactions, possibly as a dominant-negative suppressor of other transcription factors. Indeed, RBTN2 can bind to the bHLH protein TALl in uiuo (Valge-Archer et al., 1994; Wadman et al., 1994), suggesting that ectopic expression of either factor may influence similar T-cell developmental pathways. Normally present in high levels in the central nervous system, in developmentally and segmentally regulated patterns (Greenberg et al., 1990), the rhombotin proteins are only minimally expressed in or absent altogether from T cells and their progenitors. Moreover, RBTN 1 reproducibly induces thymic lymphomas after the gene (under the control of the proximal Lck promoter) is expressed by the developing thymocytes of transgenic mice (McGuire et al., 1992). In these mice, inappropriate expression of a LIM-containing protein appears to have selectively transformed a rare subset of CD8+, CD4-, and CD3- thymocytes that also express heat-stable antigen. Completing the story of developmental genes inappropriately placed under the control of TCR loci is HOXl I, located on chromosome 10, band q24 (Hatano et al., 1991; Kennedy et al., 1991; Lu et al., 1991; Dube et d., 1991). This gene encodes a homeodomain transcription factor that can bind DNA and trans-activate specific target genes (Dear et al., 1993). It is most closely related to Hh,a murine homeobox gene expressed in specific hematopoietic cell lineages and during embryogenesis (Allen et al., 1991), and more distantly related to the Antennupedia homeobox genes of Drosophila. HOXl1 normally is expressed in specific regions of the branchial arches and ectoderm of the pharyngeal pouches of the developing hindbrain in the mouse, but not in developing thymocytes or resting or activated T cells (Roberts et al., 1994; Raju, 1993). Thus, activation of H O X l 1 by chromosomal translocations in T cells, either t(10; 14)(q24;qll) or t(7;10)(q35;q24), could well lead to aberrant expression patterns of target genes that have developmentally significant roles and, thus, oncogenic potential. T h e prototypic fusion gene in ALL is BCR-ABL, formed by the der(22) of t(Y;22)(q34;q1 I ) , more commonly termed the Philadelphia chromosome. Although the functional consequences of this rearrangement are increased activity of the ABL tyrosine kinase and relocation of this enzyme from the nucleus to the cytoplasm (Kelliher et al., 1991; Van Etten et al., 1989), cloning of the Philadelphia chromosome break point opened the way for the discovery of additional chimeric oncoproteins that have often involved transcription factors. T h e E2A-PBXI chimera, resulting from a t( 1 ; 19)(q23;p13) chromosomal translocation that af-

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

33

fects about one-fourth of pre-B-cell ALLs, is perhaps the best known. In this rearrangement, the E 2 A gene on chromosome 19, which encodes a bHLH transcription factor, is fused to a homeobox gene ( P B X l ) on chromosome 1, leading to the production of several different forms of hybrid EPA-PBX1 oncoproteins (Kamps et al., 1990; Nourse et al., 1990; Izraeli et al., 1992; Numata et al., 1993). In contrast to the products of other bHLH genes affected by chromosomal rearrangements, these hybrid proteins retain only the amino-terminal trans-activation domain of E2A. T h e bHLH domain is absent, replaced by the homeobox DNAbinding and protein interaction domains of PBXl. This suggests that the gene targets of the chimeric E2A-PBX1 protein are recognized by the homeodomain of PBXl , whose normal tissue distribution does not include lymphoid o r other hematopoietic cells. In agreement with this interpretation, studies have demonstrated that PBXl is a sequencespecific DNA-binding protein, which is converted into a positive tran!scriptional regulator upon acquisition of the trans-activation domain of E2A (Van Dijk et al., 1993; Lu et al., 1994; LeBrun and Cleary, 1994). T h e transforming capacity of leukemic transcription factors generally is quite specific for particular types of human and murine hematopoietic cells. T h e E2A-PBX1 protein appears to violate this principle because, although its involvement in human disease is restricted to childhood ALL with a pre-B-cell phenotype, studies of Kamps and Baltimore (1993) have demonstrated the rapid induction of AML in lethally irradiated mice repopulated with bone marrow stem cells that were infected with recombinant retroviruses containing E2A-PBXl fusion genes. Still further heterogeneity is represented by the induction of thymic lymphomas in transgenic mice harboring E2A-PBXl genes in the germ line (Dedera et al., 1993). In these transgenic mice, lymphopenia involving both T and B cells preceded malignant transformation, suggesting that E2A-PBX1 proteins can induce apoptosis in murine lymphocyte precursors. Thus, failure to induce pre-B leukemias in these experimental systems may reflect a heightened sensitivity of the murine lymphoid compartment of EPA-PBX l-induced, programmed cell death. Of further interest is that transformation mediated by E2A-PBXl in the murine system not only depends on trans-activation motifs contributed by E2A but also (and unexpectedly) does not require the PBXl homeodomain, suggesting that E2A-PBX1 does not need to bind to DNA directly to transform cells (Monica et al., 1994). These findings suggest an alternative model in which E2A-PBXI interacts with other proteins, possibly those encoded by the major HOX loci (Van Dijk and Murre, 1994; Chan et al., 1994), to positively alter gene expression, leading to leukemogenesis.

34

A. THOMAS LOOK

The vast majority of human ALLs with E2A-PBXI gene fusions originate in B-cell precursors with cytoplasmic but not surface expression of Ig heavy chains, a characteristic of the so-called pre-B-cell phenotype. A recurring question is whether t( 1 ;19) has the same molecular genetic repercussions when it occurs in leukemias with different phenotypes. Using the reverse polymerase chain reaction (PCR) to amplify junctional sequences from leukemic cell RNA, we therefore analyzed 17 cases of t( 1;19)-positiveALLs (Privitera et al., 1992). Typical E2A-PBX1 chimeric transcripts were found in 10 of the 11 pre-B-cell cases expressing cytoplasmic Ig, but not in the six with a less mature pro-B-cell phenotype; the latter cases also lacked evidence of E2A-PBXl fusion at the genomic DNA level. These findings suggest that cytogenetically identical t( 1;19) translocations can affect entirely different loci on chromosomes 1 and 19, depending on the developmental status of the B-cell progenitor. A second fusion gene in ALL that incorporates elements of E 2 A is produced by t( 17; 19)(q22;p13), which joins N-terminal sequences of E2A (including the trans-activation domain) to the DNA-binding and protein dimerization regions of a previously unidentified hepatic leukemia factor gene (HLF) within the bZIP family of transcription factors (Fig. 2) (Inaba et al., 1992; Hunger et al., 1992). HLF, which is expressed predominantly in the liver, brain, and kidney but not normally in lymphoid cells, bears significant homology to DBP (Mueller et al., 1990), an albumin gene promoter D box-binding protein, and to TEF (thyrotroph embryonic factor), which trans-activates thyroid-stimulating hormone-@ expression during anterior pituitary development (Drolet et al., 199 1). Consistent features of E2A-HLF-associated leukemias include a proA-cell immunophenotype, onset of the leukemia in early adolescence, hypercalcemia and disseminated intravascular coagulation at diagnosis, and a poor prognosis even with intensive multiagent ALL chemotherapy. Although cases with E2A-HLF chimeric genes are quite rare, additional examples have been identified, including two in which alternative E2A-HLF fusion proteins were detected (Hunger et al., 1994b). What is the role of E2A-HLF fusion proteins in pro-B cells? Do they function primarily as DNA-binding transcription factors or in proteinprotein interactions? Findings in my laboratory (Inaba et al., 1994) and others (Hunger et al., 1994a) have identified a 10-bp DNA consensus sequence, containing a bZIP-related core dyad-symmetric motif, that specifically mediates the binding of both HLF and E2A-HLF proteins in ~lztro.Further, the chimeric protein appears to bind preferentially as a homodimer in leukemic cells, suggesting that its transcriptional regulatory effects may not require cross-dimerization with other bZIPs o r more divergent proteins (Inaba et al., 1994). T h e oncogenic potential of

Normal Chromosome 19

-

e

E2A gene

Normal Chromosome 17

Chromosome with 17;19 translocation E2A-HLF fusion gene

Chimeric mRNA Interacts with

.............. ....,,,,,,.,............................................. .......................................... ...............

1 v

Normal

Poly A sequence

Chimeric Transcription Factor

bHLH DNA binding regi

- DNA

bZIP domain

* *

- DNA

DNA

Normal

I

Interacts

-

Altered Transcription 7 Pro-B Cell Acute Lymphoblastic Leukemia 3 FIG. 2. Schematic diagram of the recently discovered E2A-HLF hybrid transcription factor. This protein, a result of t(17;19) translocation in pro-B lyrnphoblasts, combines the trans-activation domain of the E2A protein with the basic region-leucine zipper (bZIP) DNA-binding and dimerization domain of HLF, a member of the bZIP family that normally regulates gene expression in hepatocytes and brain and renal cells. The chimera may bind to DNA sequences normally recognized by the HLF protein in the liver, brain, and kidney or perhaps competes with a close homologue for a vital developmental gene in pro-B lymphoid cells. The effector (or trans-activator) region appears to function in the same manner as it does in its normal context as part ofthe E2A protein. Thus, downstream responder genes may be dysregulated, leading to the development of ALL.

36

A. THOMAS LOOK

E2A-HLF was recently established in murine NlH-3T3 cells, where the fusion protein induced anchorage-independent growth and rendered the cells tumorigenic in nude mice (Yoshihara et al., 1994). Proteins lacking the transactivation domain of E2A o r the leucine zipper dimerization domain of HLF were inactive, demonstrating a requirement for both elements in cell transformation. Taken together, these findings suggest a model in which homodimeric E2A-HLF DNA-binding complexes positively subvert transcriptional programs that normally are quiescent or actively repressed during lymphoid cell development. Whether the gene targets of E2A-HLF-mediated transformation contain binding sites recognized by HLF in liver and kidney or similar sites ordinarily bound by transcription factors in pro-B cells awaits clarification. A remarkably diverse group of chromosomal translocations and deletions affects the q23 band of chromosome 11, accounting for as many as 10% of acute leukemia cases, both lymphoid and myeloid, in children and adults (Mitelman, 1991). Of the 20 or more different chromosomal loci that have been identified as fusion partners in 1lq23 translocations, the most frequent resides on chromosome 4q21. In childhood ALL, approximately 4% of the cases overafl and a much higher percentage of infant cases have 1 lq23 translocations, one-third of which involve chromosome 4 (Raimondi, 1993). Leukemic blasts carrying t(4; 1 l)(q21;q23) have a pro-B-cell phenotype, with expression of HLA-DR antigens and rearranged Ig heavy-chain genes (Crist et al., 1985; Mirro et al., 1986; Nagasaka et al., 1983; Stong et al., 1985). There are also several distinctive clinical features associated with this rearrangement, e.g., leukocyte counts of >lo0 x 10" per liter and an adverse prognosis even with intensive combination chemotherapy (Arthur et al., 1982; Bloomfield et al., 1986; Rivera et al., 1991). Unlike most other chromosomal break points in acute leukemia, rearrangements of the l l q 2 3 region can be found in either lymphoid or myeloid malignancies. Examples are the high frequency of a t(9; 1l)(p22;q23) translocation in acute monocytic leukemia [AMoL; M5 subtype in the French-American-British (FAB) classification scheme] (Diaz et al., 1986) and of this and other llq23 abnormalities in cases with mixed lymphoid-myeloid features or in myeloid leukemias arising after treatment for ALL (Pui et al., 1989). In addition, leukemic lymphoblasts with 1lq23 abnormalities can be induced to express monocytic features in vitro, suggesting that the transformed progenitors have the potential to develop in either the lymphoid or myeloid pathway (Nagasaka et al., 1983; Stong et al., 1985). Regardless of the chromosomal lesion intruding in the llq23 region or the phenotype of the resulting acute leukemia, the genetic consequences at this locus are essentially the same: the MLL gene (also called H R X , ALL-1,

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

37

and HTRX) is disrupted within a small region situated midway through the coding region of the gene (Ziemin-van der Poel et al., 1991; Tkachuk et al., 1992; Gu et al., 1992; Djabali et al., 1992; Domer et al., 1993; Morrissey et al., 1993). Insights into the mechanisms by which MLL might contribute to leukemia have been gained by examination of its primary structure. T h e gene encodes a large (430-kDa) protein with several different structural miotifs [three N-terminal A-T hooks typical of minor groove DNAbinding proteins, a so-called M T region with homology to mammalian mlethyltransferases, two central zinc finger regions, and several regions showing homology to the Drosophila trithorax gene (Morrissey et al., 1!393; Tkachuk et al., 1992; Gu et al., 1992; Domer et al., 1993)l. When affected by rearrangements, the resulting chimeric protein always includes the MLL amino terminus; thus, it loses its zinc finger regions while retaining the A-T hook and M T domains (Fig. 3). This suggests a 1000 A-T hooks

2000

3000

I

I

I MT

Zn-fingers

TRX homology

MLL A-T hooks

MT

MLL-AF4

Fusion point SP-rich

4

4;11 der(l1) product A-T hooks

Fusion point

MT

1SP-rich

MLL-ENL

11;19 der(l1) product

Homology region with the Drosopbila Trithorax Homeotic gene regulator

1 A-T hook DNA binding motif 1Methyl transferase homology region Zinc-finger regions

FIG. 3. Important fusion proteins resulting from chromosomal rearrangements that afFect chromosome 1 1 , band q23. t(4; 1 l), which gives rise to the MLL-AF4 gene, is closely associated with ALL in infants younger than 1 year, while t(11;19) and its fusion gene, MLL-ENL are found in both ALL and AML. A characteristic feature of fusion genes involving loci in the 1lq23 region is loss of the MLL zinc finger domains, with retention of the A-T hook domains and a region homologous to mammalian methyltransferases (MT). Break points in the MLL gene are similar whether the fusion product is associated with AlML or ALL, indicating that disruption of MLL function is important in development along either lymphoid or myeloid pathways.

38

A. THOMAS LOOK

common mechanism of tumorigenicity in which A-T hooks, which may bind to DNA in the minor groove, would allow interactions with potential target gene promoters and access to other transcription factors, while the M T domain could affect transcription by modifying the DNA methylation status. Thus, the identity of the fusion partner in these varied abnormalities may specify a lineage predilection for the resulting leukemia, but be of secondary importance to retention of the A-T hook and M T domains of MLL in the oncogenic chimera. Appropriate model systems are needed to explore the full mechanistic implications of these multidomain fusion proteins. The MLL gene may be a more common participant in leukemogenesis than has been appreciated from the study of 1lq23 cytogenetic changes. Schichman and co-workers (Schichman et al., 1994a,b) discovered that MLL can be rearranged by tandem duplication of an internal portion of the gene, in the absence of l l q 2 3 translocations. This process results in linkage of the intact gene to a duplication of its amino-terminal region. Among the cases reported to date, all partial duplications have contained the A-T hook and DNA methyltransferase motifs, reinforcing the notion that these regions have critical importance in leukemia induction. B. ACLTEMYELOIDLEUKEMIAS

In AML, the EVIl gene provides a prime example of a transcription factor whose expression pattern is altered by juxtaposing promoterenhancer sequences from one chromosome to an oncogenic transcription factor on another chromosome. This dysregulation occurs by chromosomal translocations and inversions involving the E V I l locus on the long arm of chromosome 3 (Morishita et al., 1992); the same effect is produced in murine myeloid leukemias by insertional mutagenesis (Morishita et al., 1988). The EVI 1 protein binds to promoter-enhancer sequences containing the GATA sequence motif and may act by interfering with regulatory signals normally mediated by the GATA family of hematopoietic transcriptional regulators (Delwel et al., 1993; Perkins et al., 1991; Funabiki et al., 1994; Kreider et al., 1993). T h e normal function of E V l l is unknown, although its tissue distribution (oocytes and kidney cells) and its dominant interfering effect on normal myelopoiesis would suggest an important developmental role in regulatory pathways that interface between proliferation and differentiation. An example of a chimeric transcription factor with oncogenic potential is the AML1-ETO protein of AML associated with t(8;21) (Miyoshi et al., 1991; Gao et al., 1991; Erickson et nl., 1992). This translocation is

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

39

the most frequent chromosomal abnormality in the myeloid leukemias of both children and adults and is most often found in myeloblasts with evidence of granulocytic differentiation (M2 designation by the FAB classification scheme). The AMLl gene contains a large domain with 68% sequence homology to the Drosophila pair-rule segmentation gene runt, which mediates sequence-specific DNA binding and provides a protein-protein interaction domain (Meyers et at., 1993). Also characterized as PEBP2al (Ogawa et al., 1993b), AMLl forms complexes with members of the CBF protein family, which lack DNA-binding domains (Wang et al., 1993; Ogawa et al., 1993a). Interestingly, one of the proteins in this family, CBFP, is involved in another major chromosomal rearrangement in AML, inversion 16, found in association with myelomonocytic differentiation and increased bone marrow eosinophils (M4-Eo designation in the FAB scheme). This inversion joins the aminoterminal sequences of the CBFP gene to the carboxyl terminus of the smooth muscle myosin heavy-chain gene (MYHI 1 ), resulting in the forrnation of a CBFP-MYH1 1 fusion protein (Liu et d., 1993). Both AMLl and CBFP appear to be normally expressed, at least in some of their spliced variant forms, in early myeloid cells, suggesting that their oncogenicity stems from disruption of a transcriptional regulatory complex specific for myeloid developmental processes (Nuchprayoon et al., 1994). T h e oncogenic versatility of the AMLZ locus can be seen in findings in which the t(3;2 1) translocation caused AMLl sequences to become linked to sequences from either the EVIl gene [in chronic myeloid leukemia in blast crisis (Mitani et al., 1994)] or the Epstein-Barr virus RNAassociated protein (EAF) gene (in myelodysplastic syndrome) (Nucifora rt al., 1993). T h e mechanistic implications of the AMLl -EVII fusion are noteworthy, in that both the runt homologous DNA-bindingdimerization domain of AMLl and the zinc finger DNA-binding domains of EVZZ are included in the chimeric protein, opening numerous opportunities for aberrant regulation of target genes whose promoters contain binding sites for either factor. The AMLl-EAP chimeric protein contains carboxyl-terminal EAP sequences fused out of frame to AMLl sequences, resulting in a truncated AMLl protein that may interfere with normal AMLl function during myelopoiesis. T h e linkage between chimeric and dysregulated transcription factors and differentiation arrest at specific stages of development in myeloid leukemia has provided a new class of intracellular targets for therapeutic atempts to differentiate these leukemias in uivo, such that they lose their proliferative capacity. A major example is the transcription factor fusion (due to t( 15;17)(q21;ql1-22) in acute promyelocytic leukemia (APML),

4u

A . THOMAS LOOK

which links critical ligand- and DNA-binding sequences of the retinoic acid-a-receptor gene (RARa) on chromosome 17 to sequences of the P M L gene on chromosome 15 (de T h e et al., 1990, 1991; Borrow et al., 1990; Longo et al., 1990; Kakizuka et af., 1991). T h e RARa nuclear receptor protein binds the retinoic acid ligand and then DNA through a rinc finger region. PML proteins, which possess potential zinc finger motifs, are normally located in novel macromolecular nuclear organelles, called PML oncogenic domains (PODS),that include at least three other proteins (Dyck et al., 1994; Weis et al., 1994; Koken et al., 1994). PML-RARa fusion proteins disrupt these subnuclear structures, dispersing normal PML, RXR, and other nuclear proteins in an abnormal microparticulate nuclear pattern (Weis et al., 1994; Dyck et al., 1994; Koken et al., 1994). T h e fusion proteins interfere with normal myeloid cell development, possibly through adverse effects on the assembly of the PODS containing PML, leading to arrested differentiation in the promyelocyte stage. These findings provide the mechanistic rationale for the use of all-trans-retinoic acid to treat patients with APML (Huang et ul., 1988; Chen et al., 1991; Warrell et al., 1991, 1994; Fenaux et al., 1994). In response to pharmacological doses of this compound, PML and its associated proteins are reorganized into normal-appearing nuclear PODS, with subsequent development of the leukemic cells into mature myeloid cells with limited life spans in the circulation. Resistance to all-tmn~-retinoicacid generally develops within 3-4 months, limiting this hormone’s therapeutic role in the remission induction phase of APML therapy, as an adjunct to cytotoxic chemotherapy (Warrell et al., 1994; Fenaux et al., 1994). C:. SARCOMAS

In addition to the leukemias, chimeric transcription factors resulting from chromosomal translocations have been identified in a large number of soft-tissue sarcomas (Table 11). T h e predilection of gene fusion events for the seemingly disparate hematopoietic and mesenchymal cell types, rather than for epidermal malignancies such as the carcinomas, for example, probably relates to the reliance of both hematopoietic and mesenchymal tissues on transcription factor networks to program the differentiation of many different cell lineages for shared stem cells and progenitors in both embryologic and postnatal development. T h e first sarcoma translocation to be characterized at the molecular level is t( 1 1 ;22), which has long been recognized as virtually pathognomonic of Ewing’s sarcoma (EWS) o r its close relative, primitive neuroec-

TABLE I1 TRANSCRIPTION FACTOR GENESAFFECTED BY CHROMOSOMAL BREAKPOINTSIN HUMAN SARCOMAS Familya

Translocation

Affected gene

E WS-ATFl FUS-CHOPb

Delattre et al. (1992) Sorensen e tal. (1994a) Jeon et al. (1995) Zucman et al. (1993a) Crozat et al. (1993); Rabbitts et al. (1 993)

t(2;13)(q35;q14)

PAX3-FKHRb

Alveolar rhabdomyosarcorna

t(1; 13)(p36;q14)

PAX7-FKHR

Alveolar rhabdomyosarcoma

Shapiro et al. (1993); Barr et al. (1993) Davis et al. (1994)

t(l1;22)(~13;q12)

E WS- W T l

Desmoplastic round cell tumor

Ladanyi and Gerald (1994)

SYT-SSXb

Synovial sarcoma

Clark et al. (1994)

EWS-FLII b EWS-ERGb EWS-ETVl

Basic region-leucine zipper (bZip) proteins Paired box/homeodomain

Zinc finger proteins Undefined

t(X; 18)(pI1.2;q11.2)

b

Based on DNA-binding domain. Fusion gene.

Reference

Ewing’s sarcoma Ewing’s sarcoma Ewing’s sarcoma Melanoma of soft parts Liposarcoma

t(l1;22)(q24;q12) t(2 1;22)(q22;q12) t(7;22)(p22;q22) t( 12;22)(q13;q12) t( 12; 16)(ql3;p 11)

Ets homology domain proteins

Disease

42

A. THOMAS LOOK

todermal tumor (PNET). The chimeric gene produced by this translocation, EWS-FLZI (Delattre et al., 1992), results in the expression of a fusion protein containing amino-terminal sequences of EWS linked to the Ets-like DXA-binding domain of FLI 1 (named for Friend leukemia integration site 1) (Ben-David et al., 1991). In uitro transformation assays were used to establish that both the EWS domain and the DNA-binding domain of the protein are required for malignant conversion (May et al., 1993a), suggesting that the chimeric protein acts by disrupting transcriptional regulatory pathways. Furthermore, variant translocations have been identified that fuse identical EWS sequences to the DNAbinding domains of two related Ets family members, ERG (Zucman et al., 1993b; Sorensen ct al., 1994b; Delattre et a!., 1994; Giovannini et al., 1994) and E T V 1 (Jeon et d.,1995), strengthening the critical association between these two regions. T h e RNA recognition motif of EWS is absent from the various oncogenic fusion proteins, replaced by the DNAbinding domains of the partner proteins in these fusions. T h e aminoterminal sequences of EWS that are included in fusion proteins are rich in glutamine, serine, and tyrosine, and this domain is a potent transactivator of gene expression (May et a.l., 1993b; Bailly et al., 1994), suggesting that fusion genes containing an EWS trans-activation domain and an Ets-like DNA-binding domain act by dominantly and aberrantly upregulating the expression of critical target genes in this sarcoma. More recently, extreme 5’ domains of EWS have been shown to possess potent transforming activity when linked to FLI 1, but to mediate little trans-activation in model systems (Lessnick et al., 1995). Thus, this region may be important in mediating protein-protein interactions that indirectly modify transcription. When the same sequences of EWS are fused to the bZIP domain of ATF1 in t(12;2), the resultant tumor is malignant melanoma of the soft parts (Zucman et al., 1993a). A different t(12;16) translocation in malignant liposarcoma fuses the amino terminus of a protein called FUS or TLS to the bZIP domain of CHOP, which was initially characterized as a non-DN A-binding dominant-negative inhibitor of other bZIP proteins of the CAAT box-binding, C/EBP family (Crozat et al., 1993; Rabbitts et ul., 1993). Both EWS and FUS/TLS are abundant nuclear RNA-binding proteins that associate in z~iuowith products of RNA polymerase IImediated transcription, in the form of ternary complexes with heterogeneous RNA-binding proteins such as A1 and C1/C2 (Zinszer et al., 1994). Furthermore, the amino-terminal sequences of FUS that are found in FUS-CHOP chimeras are experimentally interchangeable with the analogous sequences from EWS in trans-activation and transformation assays (Zinszer ef al., 1994), supporting the concept that the DNA-

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

43

binding region of the fusion protein specifies the activated downstream target genes and, thus, the phenotype of the arrested and transformed malignant mesenchymal progenitors. This principle is well-illustrated by the interesting EWS-WTl fusion gene, in which EWS sequences are fused to the zinc finger DNA-binding domain of the Wilms’ tumor gene, W T I , in desmoplastic round cell tumors (Ladanyi and Gerald, 1994). WTl is an important tumor suppressor gene when its expression is horriozygously inactivated in Wilms’ tumor, emphasizing the versatility of transcription factor involvement in embryonal tumors affecting different cell lineages. In alveolar rhabdomyosarcoma, the DNA-binding regions of two different PAX proteins, PAX3 and PAX7, form fusion proteins with a portion of a forkhead domain gene, called FKHR (Shapiro et al., 1993; Barr el al., 1993; Davis et al., 1994). The PAX genes are important developmental regulators (see the following) that bind DNA and presumably recognize specific target genes through their paired box and homeodomain DNA-binding regions. The t(2; 13)(q35;q14) rearrangement preserves both DNA-binding domains (paired box and homeodomain) of the PAX3 gene, while replacing its carboxyl-terminal sequences with bisected forkhead DNA-binding sequences from FKHR. A variant translocation, t(1;13)(p36;q14), has a similar effect on PAX7 (Davis et al., 1994), suggesting that the two genes may regulate a common set of target genes involved in the pathogenesis of rhabdomyosarcoma. Finally, the SYT-SSX gene of synovial sarcoma remains poorly characterized because neither fusion element shows homology to previously described sequences (Clark et al., 1994).

111. Oncogenic Transcription Factors and the Developmental Regulatory Proteins of Drosophila A. HOMOLOGIES WITHIN DNA-BINDING AND DIMERIZATION DOMAINS Intriguing structural similarities have emerged between the transcript ion factor genes discovered at chromosomal break points and regula-

tory proteins linked to the control of Drosophila development. As shown in Table 111, the regions of homology between transcription factors implicated in the leukemias and sarcomas and those with roles in Drosophila segmentation are most striking in the DNA-binding and dimerization domains, which specify the DNA sequence elements identified by transcriptional regulators and, in many cases, provide interfaces for protein-protein interactions. Drosophila segmentation genes are classi-

44

A. THOMAS LOOK

TABLE 111 EVOLLTlON4RILY <:ONSERVED S T R U C T U R A L MOTIES LINKING OXCOGENIC TRANSCRIPTION FACTORS wn H PROTEINS REGULATINGSEGMEXTATION PATTERNIN Drvsvphzla ~~

~

~

Cancer type.

Human protein

DNA-binding domain

Morphogenetic role

Drvsvphrla protein

~~~

bZIP

Giant

Gdp

Zinc finger

Kriippel

Gap

Runt homology

Runt

Pair rule

Paired box

Paired

Pair rule

Alveolar RMS

H LF PLZF AML1 PAX3 FKHR

Forkhead

Sloppy paired

Pair rule

T ALI. pre-B .4LL

HOXl1 PBX 1

Hoineobox Homeobox

Antennapedia Extradentricle

Homeotic Homeotic

pro-B .4LL/.4ML

MLLIHRXIALL 1

A-T hook

Trithorax

Hoineotic gene regulator

Pro-B .4LL APML

AMI. Alveolar RMS

Abbreviations: ALL, acute lvmphoblastic leukemia: AMI., acute myeloid leukemia; APML, acute proin!eloc)tir leukemia; RMS, rhabdomyosarcorna. 0

Maternal polarity genes

-

genes

---

Kruppel (Kr)

* Giant (gt)

Knirps (Kni)

-

Hunchback (hb)

Tail-less (tll) Huckebein (hkb)

= Bicoid (bcd)

genes

-

Pair-rule genes

Nanos (nos)

Antennapedia Complex (antp)

= Bithorax Complex (ubx) Extradenticle (exd)

egment polarity genes

-

---

Hairy (h) Even-skipped (eve) Aunt (run)

Fushi tarazu (ftz) Paired (prd) Sloppy-patred (slp)

Engrailed (en) Wingless (wg) Patched (ptc) Hedgehog (hh)

FIG.4. Developmental gene regulation during segmentation in Drvsvphilu. Each step in this cascade is controlled by a discrete family of nuclear regulatory genes whose sequences have been conserved in mammalian oncogenic transcription factors active in the acute leukemias and sarcomas (see Table 111).

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

45

fied according to a model developed over the past two decades, based on the molecular control of pattern formation during the embryogenesis of this organism (Nusslein-Volhard and Wieschaus, 1980; NussleinVolhard et ul., 1987; Levine and Harding, 1989). Thus, as illustrated in Fig. 4, gap genes are among the first to be transcribed in the developing embryo and are expressed in broad domains, approximately three segments wide, in response to gradients of cytoplasmic polarity proteins derived from maternal mRNAs. The term “gap” refers to the characteristic spaces produced in segmentation patterns of fly embryos harboring mutations of these genes. Two of the gap proteins-giant and Kriippel-have mammalian counterparts in the HLF (bZIP) and PLZF (zinc finger) proteins of pro-B-cell ALL and acute promyelocytic leukemia, respectively. Gap proteins act in concert to regulate the expression of pair-rule genes, which are transcribed in the primordia of every other segment, form repetitive segmental divisions in the embryo, and are expressed in a pattern of seven stripes along the anterior to posterior axis. Mutations of pair-rule genes in Drosophila generally result in the deletion of portions of every other segment. T h e mammalian AMLl protein, associated with the t(8;2 1)chromosomal translocation often found in acute myeloid leukemias, shares homology with a major pair-rule protein (runt) (Erickson et al., 1992), as do the PAX3 and FKHR proteins fused by the t(2; 13) rearrangement in alveolar rhabdomyosarcoma (paired and sloppy paired, respectively) (Gruss and Walther, 1992; Grossniklaus et al., 1992). Pairrule proteins regulate the expression of segment polarity genes, which are responsible for the formation of certain repeated structures, such as the boundary regions, common to each Drosophilu segment. Mutations of these genes cause the loss of a portion of each segment, which is replaced by a mirror image structure from the same segment. Segment polarity, pair-rule, and gap proteins interact with other regulatory proteins to control the expression of homeotic genes, whose products determine the unique structures (e.g., antennae, wings, legs) expressed by each segment. As shown in Table 111, both the HOXl 1 and PBX 1 homeoproteins have counterparts in the homeotic class of segmentation proteins, the latter having extensive homology that is not lirnited to the homeodomain itself (Rauskolb et al., 1993). Both Drosophilu ex tradenticle and human PBXl have been shown to physically associate arid act in concert with the major homeotic selector proteins in their respective organisms, presumably to modulate morphological manifestations during development (Van Dijk and Murre, 1994; Chan et al., 1994). T h e MLL protein, involved in at least 20 specific translocations, deletions, inversions, or tandem duplications in both lymphoid and my-

46

A.

THOMAS LOOK

eloid leukemias, is the mammalian counterpart of the trithorax protein, also a Drosophila homeotic gene regulator (Mazo et al., 1990; Djabali et al., 1992; Tkachuk et al., 1992; Gu et al., 1992; Domer et al., 1993). Thus, distinct classes of Drosophila genes encode transcriptional regulatory proteins whose patterns of expression are determined by a hierarchical cascade in the developing embryo (Fig. 4). Ultimately, these transcription factor genes regulate the developmental programs of structural genes that define the three-dimensional form of the unfolding embryo. Structural similarities between the products of these genes and the oncogenic transcription factors identified in many of the leukemias and sarcomas are important because they suggest that the normal counterparts of the oncogenic proteins may have important developmental roles in mammalian embryogenesis. TRANSCRIPTION FACTORS IN B. ROLESOF ONCOCENIC NORMAL MAMMALIAN DEVELOPMENT

T h e hypothesis that structurally similar transcription factors in Drosophila and humans share normal functional roles during development has been explored by producing loss-of-function mutations in specific genes of mouse embryonic stem cells and examining the consequences of these manipulations during mammalian developtnent in vivo. Although saturating mutagenesis was employed to identify genes important in Drosophilu embryogenesis, the types of recessive, homozygousinactivating mutations that were produced nonetheless are directly analogous to those produced by targeting individual genes in murine systems. Both approaches can identify the earliest function of a gene product during development, but later effects may be obscured if the gene is essential for embryonic viability. Iniportant functions may also be inapparent when other gene products have similar or overlapping functions (so-called “redundancy”), especially under the more complex conditions required for the study of mammalian development. Despite experimental limitations, the morphogenetic roles of several transcription factor genes identified at chromosomal break points are becoming clear. One example is the HOXll gene, which is activated by translocation in certain T-cell acute leukemias. HOXll is classified as an “orphan” homeobox gene, because it resides outside of the linear arrays of major homeobox proteins in the genome, which encode proteins that act in concert to regulate segment-specific gene expression along the anterior-posterior axis of the embryo (McCinnis and Krumlauf, 1992). A very specific homeotic role of HOXll in mammalian development was indicated when homozygous disruptions of this gene were found to

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

47

block the formation of the spleen in otherwise normal mice (Roberts et al., 1994). H O X l l is expressed in structures arising from the first three branchial arches and the hindbrain at early stages of development, as well as from a single site corresponding to the splanchnic mesoderm beginning at embryonic day 11.5 (Roberts et al., 1994). Thus, the roles of hrOXl1 proteins in branchial arch and hindbrain structures appear to be developmentally insignificant or at least compensated for by other transcription factors expressed by these cells; however, their role in cellular organization at the site of splenic development is absolutely essential for the genesis of this organ. Interestingly, lymphoid and other types of hematopoietic cells, which normally lack H O X l l proteins, were not affected by loss-of-function mutations in this gene, except for the presence of asplenia-related Howell-Jolly bodies in circulating erythrocytes. The mammalian PAX family, which shares paired-box DNA-binding domains with several Drosophilu segmentation genes, comprises at least eight members, two of which, PAX? and PAX7, have been identified in fusions with the FKHR gene in the transformed skeletal muscle progenitors of alveolar rhabdomyosarcoma (Shapiro et al., 1993; Barr et al., 1993; Davis et al., 1994). Studies of the chromosomal location and pattern of PAX expression during murine development led Gruss and colleagues to link partial or complete disruption of genes to several previously identified inherited disorders [reviewed in Gruss and Walther (1992)]. I n one mutant mouse phenotype, designated undulated, PAXl shows a point mutation leading to a single amino acid substitution in the paired-box domain. The affected mice have defects of the axial skeleton, consistent with restriction of the PAXl expression pattern of mesodermal derivatives. Alterations of PAX?, by contrast, lead to a mutant murine phenotype called splotch, which in its homozygous form is characterized by abnormalities of the central nervous system and structures arising from the neural crest, including exencephalus, spina bifida, dysgenesis of the spinal ganglia, and abnormalities of pigmentation. Of particular interest, mutations of the human PAX? gene cause an inherited human syndrome called Waardenburg syndrome I, an autosomal dominant condition characterized by a depigmented forelock, lateral displacement of the inner ocular canthus, deafness, and mental retardation. Apparently tlhe PAX3 product not only is essential during developmental processes but also is required in high concentrations, as the dominant mode of inheritance of Waardenburg syndrome I would seem to imply that disruption or mutation of only one allele causes levels of this protein to fall below a critical threshold, leading to phenotypic consequences. A consistent theme in the study of these naturally occurring or artificially induced loss-of-function mutations is the similarity they reveal between

48

A. THOMAS LOOK

the normal roles of master regulatory proteins in Drosophila and mammalian embryological development. Unfortunately, the insights provided do not directly explain the oncogenic activation of these transcription factors in human leukemias and sarcomas, which generally involves gain-of-function molecular changes in cells that normally do not express the affected genes. However, the parallels with normal developmental control suggest that these oncogenic mechanisms involve perturbations of networks of regulatory proteins that have evolved from more primitive embryological pathways to meet the needs of specialized mammalian hematopoietic and mesenchymal systems.

C. ONGOGENIC TRANSCRIPTION FACTORSMAY DISRUPT NORMAL REGULATORY NETWORKS THAT DETERMINE HEMATOPOIETIC A N D MESENCHYMAL CEIL FATE How might dysregulated or chimeric transcription factors contribute to the malignant transformation of so many different types of blood cell and soft-tissue progenitors? Conceivably, chromosomal rearrangements in susceptible progenitor cells with proliferative capacity could initiate the malignant process, or they could occur later as a result of mutations that impart genomic instability. Whatever the sequence, the critical event seems to be the abnormal production of a functional protein from one copy of the translocated gene(s), which by interaction with DNA or other proteins is able to create differentiation arrest and growth disturbances in a particular cell lineage, leading to the selection and proliferation of the affected cells as a malignant clone. It would be tempting to propose a tumorigenic model based on general regulators of the cell cycle; however, each hybrid o r dysregulated factor is so tightly linked to a given phenotype of leukemia or sarcoma (Tables I and 11) that it must disrupt unique aspects of developmental regulation that are restricted to that lineage. Thus, the most reasonable hypothesis would be that these proteins, as descendants of analogous factors controlling Drosophila embryogenesis, probably interfere with the normal function of regulatory cascades specific for different stages of cell development. For example, research in my laboratory suggests that the oncogenic E2A-HLF transcription factor competes for key target genes with other factors normally expressed in early B lymphoid progenitors (Fig. 5 ) (Inaba et d., 1994). This evidence does not, however, eliminate models in which the critical oncogenic event is inappropriate expression of otherwise silent genes, such as those recognized by HLF in the liver. Alternatively, some regulatory proteins, such as those in the

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

49

A Leukemlc

Normal

E4BP4

E4BP4

Target Gene

B Normal

Leukemlc DBPlEPAHLF

Downstream Targets (Possibly other transcription factor aenes)

DNA

Target Gene

Target Gene

FIG. 5. Competitive models to account for the oncogenic role of the E2A-HLF chimeric protein in pro-B-cell ALL. Aberrant expression of master transcription factors in mammalian blood cells is thought to subvert genetic programs controlling important steps in hematopoiesis. Study of the E2A-HLF chimera suggests a mechanism of leukemia induction in which the altered protein competes as a homodimer (A) or a heterodimer (B) with transcription factors normally expressed in pro-B lymphoblasts, either E4BP4 or DBP, for example. In these models, E2A-HLF trans-activated the expression of a normally repressed gene, leading to inappropriate activation of downstream targets (which possibly are other transcription factor genes); this in turn modified regulatory cascades that drive pro-B-cell development, leading to differentiation arrest and abnormal proliferation characteristic of leukemic cells.

LIM group, may form inactive complexes that interfere with normal regulatory proteins through a dominant-negative mechanism. Gene inactivation experiments in embryonic stem cells that target transcription factors normally involved in the regulation of lineagespecific genes are beginning to reveal the outlines of regulatory cascades in normal hematopoietic cell development. For example, lymphoid cell development has been shown to depend on the function of at least two transcription factors, the zinc finger Ikaros protein (Georgopoulos et al., 1994) and the Ets family member PU. 1 (Scott et al., 1994). Similarly, early erythroid cell development depends on the function of several proteins, including two that were first identified because of their involvement in T-cell leukemia translocations, Rbtn2 (Warren et al., 1994) and Tall (Stuart Orkin, personal communication), as well as the zinc finger pro-

50

A. THOMAS LOOK

teins Gatal (Pevny et al., 1991), Gata2 (Tsai et al., 1994), and c-Myb (Mucenski et al., 1991). It seems remarkable that two genes identified by reverse genetics at the break points of T-cell translocations, whose products have been shown to form niultimolecular complexes with each other (Valge-Archer et al., 1994; Wadman et al., 1994), should turn out to be essential for normal erythropoiesis. If shared downstream target genes are involved in both the normal and abnormal developmental processes in these two lineages, then research on the oncogenic mechanisms of these genes should converge with studies of normal red blood cell ontogeny. A similar network of regulatory factors has emerged from studies of muscle cell development [reviewed in Weintraub (1993) and Buckingham (19Y4)I. In this research, myogenesis was found to be regulated, at least in part, by master genes of the MyoD family that encode several basic helix-loop-helix proteins (MyoD, Myf-5, MRF4, and myogenin), which, like the related Tall protein, form heterocomplexes with the E12iE47 products of the E2A gene. T h e similarities between the genes involved in normal development and the genes activated by chromosomal translocation in these lineages suggest that oncogenic transcription factors will serve as probes to aid in dissecting the networks of nuclear proteins that control normal lymphopoiesis, myelopoiesis, and mesodermal cell development, while providing insight into the pathogenetic mechanisms underlying human acute leukemias and sarcomas. IV. Summary and Future Directions What has been learned about the transforming roles of transcription factors in leukemias and sarcomas? First is the requirement for protooncogene activation, often by chromosomal rearrangements (either reciprocal translocations, inversions, deletions, or tandem duplications) in which the candidate gene comes to lie in the vicinity of a TCR or Ig gene or is fused with a second gene to form a chimeric protein that retains many of the key functions of the original factors. Postactivation regulatory events remain largely unknown, although the array of factors identified so far suggests an extraordinarily diverse set of interactions. For example, heterodimerization with other proteins, as in the formation of MYC:MAX, AMLl:CBFP, TALl:RBTNP, or PBX1:HOX complexes, could increase the complexity of interactions between oncogenic proteins and transcriptional regulatory networks. On the other hand, as postulated for the EPA-HLF chimera, immediate postactivation events may be limited to the formation of homodimers, thus restricting the repertoire of available downstream target genes. They key to understanding the oncogenic effects of transcription

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

51

factors lies in the nature of the genes they regulate. Since the majority of these master oncoproteins are ectopically expressed, one might predict that they alter the expression of tightly regulated genes in normal hematopoietic o r mesenchymal progenitors. Almost certainly examples will be found in which interaction of these proteins with downstream target genes either activates or represses developmental programs that are normally required only at critical times in the life cycle of the progenitor cell. I n many instances, the positive or negative effects of oncogenic transcription factors are probably mediated directly through binding to enhancer sequences in target gene promoters; however, in other cases these proteins may transform cells indirectly by binding to other transcriptional regulatory proteins and targeting them to nonfunctional or newly functional complexes. Thus, the main challenge for the future is to identify the genes controlled by the various transcription factors activated by chromosomal rlearrangements. Equally important will be the task of delineating subsequent interactive processes within transcriptional regulatory cascades. This is likely to be even more difficult than deciphering the molecular mechanisms regulating Drosophila embryogenesis, because of the greater complexity of experimental embryology and genetic analysis in vertebrate model systems. With increased knowledge of the regulatory networks affected by oncogenic transcription factors, it will be possible to develop new therapeutic strategies for human malignancies similar to those employing rletinoic acid for the treatment of acute promyelocytic leukemia. Indeed, if’ a fusion protein is crucial to the persistence of a tumor cell’s malignant growth, one could alter the disease’s course by interfering with any of the multiple steps in protein synthesis and action, including oncogene transcription, RNA processing and translation, and DNA or protein interactions. Fused transcription factors would appear to be ideal for these types of intervention, because they represent true chimeras that occur only in rare types of malignant cells. A clear advantage of this approach, which depends on a detailed understanding of the mechanisms underlying each hybrid factor’s transforming properties, would be the reduced likelihood of toxicity to normal cells or the development of resistant mutants, when compared to currently available methods of cancer therapy. ACKNOWLEDGMENTS I thank John Gilbert for editorial review and critical comments. Supported in part by NIH Grants CA-59571, CA-21765, and CA-20180 and by the American Lebanese Syrian Associated Charities (ALSAC).

52

A . THOMAS LOOK

REFERENCES Adams, J. M . , Gerondakis, S., Webb, E., e! al. (1983). Proc. Natl. Acad. Sci. U S A 80, 1982. Adams, J. M . , Harris, A. W., Pinkert, C. A., Corcoran, L. M., Alexander, W. S., Cory, S., Palmiter, R. D., and Brinster, R. L. (1985). Nafure 318, 533-538. Allen, J. D., Lints, T., Jenkins, N . A,, Copeland, N . G., Strasser, A,, Harvey, R. P., and Adams, J. M. (1991). Genes Deu. 5, 509. Amati, B., Brooks, M. W., Levy, N.. Littlewood, T. D., Evan, G. I., a n d Land, H . (1993). Cell 72,233-245. Arthur, D. C., BLoomfield, C. D., Linquist, L. L., and Nesbit, M. E. J. (1982). Blood 59,96. Ayer, D. E., and Eisenman, R. N. (1993). Genes Dew. 7, 21 10-21 19. Ayer, D. E., Kretzner, L., and Eisenman, R. N. (1993). Cell 72, 21 1-222. Baer, R. (1993). Sen. CancerBiol. 4, 341-347. Bailly, R. A . , Bosselut, R., Zucman, J., Cormier, F., Delattre, O., Roussel, M., Thomas, G., and Ghysdael. J. (1994). Mol. Cell Biol. 14, 3230-3241. Barr, F. G., Galili, N., Holick, J., Biegel, J. A., Rovera, G., and Emanuel, B. S. (1993). Nature Gene!. 3, 1 13- 1 17. Hegley, C . G., Aplan, P. D., Davey, M. P., Nakahara, K., Tchorz, K., Kurtzberg, J., Hershfield, M .S., Haynes, B. F., Cohen, D. I . , Waldmann, T. A,, et al. (1989).Proc. Natl. Acad. Sci. L'S.4 86, 203 1-2035. Hen-David, Y., Giddens, E. B., Letwin, K., and Bernstein, A. (1991). Genes Dev. 5,908-918. Blackwood, E. M . . a n d Eisenman, R. N . (1991). Science 251, 1211-1217. Bloomfield, C. D.. Goldman, A. L., Berger, A. R., et al. (1986). Blood 67, 415. Hoehm, T., Foroni, L.. Kaneko, Y., Perutz, M. F., and Rabbitts, T. H. (1991). Proc. Nutl. Acad. Scz. C'S4 88, 4367-437 1 . Borrow, J.) Goddard, A. D., Sheer, D., and Solomon, E. (1990). Science 249, 1577-1580. Buckingham, M. (1994). Cell 78, 15-21. Chan, S., Jaffe, L.. Capovilla, M., Botas, J., and hlann, R. S. (1994). Cell 78, 603-615. (;hen, Q., Cheng, J. T., Tasi, L. H., Schneider, N., Buchanan, G., Carroll, A., Crist, W., Ozanne, B., Siciliano, M. J., and Baer, R. (1990). E M B O J . 9, 415-424. Zhang, R., Tao, R. F., Xia, X. M., Li, C., Wang, W., Zu, W. Y., Yao, Chen, Z.X., Xue, Y. Q., X. Z., and Ling. B. J. (1991). Blood 78, 1413-1419. Chen, Z., Brand, N., Chen, A,, Chen, S., Tong, J., Wang, Z., Waxman, S., and Zelent, A. (1993). E M B O J . 12, 1161-1167. Clark, J., Rocques, P. J., Crew, A. J., Gill, S., Shipley, J., Chan, A. M.,Gusterson, B. A., and Cooper. C . S. (1994). iVa/ure G m r f . 7, 502-508. Crist, W. M., Cleary. M. L., Grossi, C. E., ef nl. (1985).Blood 66, 33. Croce, C . M., Thierfelder, W., Erikson, J., Nishikura, K., Finan, J., Lenoir, G. M . , and Nowell, P. C. (1983). Pror. ,Vatad. Acad. Sri. L'SA 80, 6922-6926. Crozat, '4.. Aman, P., Mandahl, N.,and Ron, D. (1993). Nuture 363, 640-644. 1)alla-Favera. R.,Bregni, XI., Erikson, J., Patterson, D., Gallo, R. C., and Croce, C. M. (1982). Proc, .Vat(. Acad. Sci. U S A 79, 7824-7827. Davis, R. J., D'Cruz, C. M., Lovell, M. A . . Biegel, J. A,, and Barr. F. C;. (1994). CancerRes. 54,2869-2872. Dear, T. N., Sanchez-Garcia. I . , and Kabbitts. T. H. (1993). Pror. ,Vutl. Acad. Scz. U S A 90, 4-131-443.5. Dedera. D. X.. LVaIler, E. K . , LeBrun. D. P.. Sen-hfajumdar, A , , Stevens, M. E., Barsh, (;. S., and Clear), M. L. (1993). Cell 74, 833-843. I)elattre, O., Zucnian. J., Plougastel, B., Desmaze, C.. Melot, T., Peter, M . , Kovar, H., Joubert, l . , deJong, P., Rouleau, G.. e! 01. (199'2). .\'atwe 359, 162-165.

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

53

Delattre, O., Zucman, J., Melot, T., Garau, X. S., Zucker, J. M., Lenoir, G. M., Ambros, P. F., Sherr, D., Phil, D., Turc-Carel, C., Triche, T.J., Aurias, A. and Thomas, G. (1994). New Engl. J. Med. 331, 294-299. d e The, H., Chomienne, C., Lanotte, M., Degos, L., and Dejean, A. (1990). Nature 347, 558-561. d e T h e H., Lavau, C., Marchio, A., Chomienne, C., Degos, L., and Dejean, A. (1991). Cell 66,675-684. Delwel, R., Funabiki, T., Kreider, B. L., Morishita, K., and Ihle, J. N. (1993).Mol. Cell. Biol. 13,4291-4300. Diaz, M. O., Le Beau, M. M., Pitha, P., and Rowley, J. D. (1986). Science 231, 265. Djabali, M., Selleri, L., Parry, P., Bower, M., Young, B. D., and Evans, G. (1992). Nature Genet. 2, 113. D’omer,P. H., Fakharzadeh, S. S., Chen, C. S., Jockel, J., Silverman, G. A., Kersey, S. J., and Korsmeyer, S. J. (1993). Proc. Natl. Acad. Sci. USA 90, 7884-7888. Dirolet, D. W., Scully, K. M., Simmons, D. M., Wegner, M., Chu, K. T., Swanson, L. W., and Rosenfeld, M. G. (1991). Genes Dev. 5, 1739-1753. Dlube, I. D., Kamel-Reid, S., Yuan, C. C., Lu, M., Wu, X., Corpus, G., Raimondi, S. C., Crist, W. M., Carroll, A. J., Minowada, J., et al. (1991).Blood 78, 2996-3003. Dlyck, J. A., Maul, G. G., Miller, W. H., Jr., Chen, J. D., Kakizuka, A,, and Evans, R. M. (1994). Cell 76, 333-343. Emanuel, B. S., Selden, J. R., Chaganti, R. S. K., et al. (1984).Proc. Natl. Acud. Sci. USA 81, 2444. Erickson, P., Gao, J., Chang, K. S., Look, T., Whisenant, E., Rairnondi, S., Lasher, R., Trujillo, J., Rowley, J., and Drabkin, H. (1992). Blood 80, 1825-1831. Erikson, J., Nishikura, K., ar-Rushdi, A., et al. (1983). Proc. Natl. Acud. Sci. USA 80, 75817585. Renaux, P., Chastang, C., Chomienne, C., and Degos, L. (1994). Lancet 343, 1033. Finger, L. R., Harvey, R. C., Moore, R. C., Showe, L. C., and Croce, C. M. (1986). Science 234, 982-985. Funabiki, T., Kreider, B. L., and Ihle, J. N. (1994). Oncogene 9, 1575-1581. Gao, J., Erickson, P., Gardiner, K., Le Beau, M. M., Diaz, M. O., Patterson, D., Rowley, J. D., and Drabkin, H. A. (1991). Proc. Natl. A c d . Sci. USA 88, 4882-4886. Georgopoulos, K., Bigby, M., Wang, J., Molnar, A., Wu, P., Winandy, S., and Sharpe, A. (1994). Cell 79, 143-156. Giovannini, M., Biegel, J. A,, Serra, M., Wang, J., Wei, Y. H., Nycum, L., Emanuel, B. S., and Evans, G. A. (1994).J. Clin. Invest. 94, 489-496. Ckeenberg, J. M., Boehm, T., Sofroniew, M. V., Keynes, R. J., Barton, S. C., Norris, M. L., Surani, M. A,, Spillantini, M. G., and Rabbits, T. H. (1990).Nature 344, 158-160. Grossniklaus, U., Pearson, R. K., and Gehring, W. J. (1992). Genes Dev. 6, 1030-1051. Gruss, P., and Walther, C. (1992). Cell 69, 719-722. Gru, Y., Nakamura, T., Alder, H., Prasad, R., Canaani, O., Cimino, G., Croce, C. M., and Cananni, E. (1992). Cell 71, 701-708. Hatano, M., Roberts, C. W., Minden, M., Crist, W. M., and Korsmeyer, S. J. (1991). Science 253, 79-82. Hollis, G. F., Mitchell, K. F., Battey, J., et al. (1984). Nature 307, 752-755. Hsu, H., Cheng, J., Chen, Q., and Baer, R. (1991). Mol. Cell. Biol. 11, 3037-3042. Huang, M. E., Ye, Y. C., Chen, S. R., Chai, J. R., Lu, J. X., Zhoa, L., Gu, L. J., and Wang, Z. Y. (1988). Blood 72, 567-572. Hunger, S. P., Ohyashiki, K., Toyarna, K., and Cleary, M. L. (1992). Genes Dev. 6, 16081620.

54

A. THOMAS LOOK

Hunger, S. P.. Brown, R., and Cleary. M. L. (1994a). Mol. Cell. Biol. 14, 5986-5996. Hunger, S. P.. Deveraj, P. E., Foroni. L., Secker-Walker, L. M., and Cleary, M. L. (l994b). Blood 83,2261-2267. . 2865-2868. Ichikawa, H., Shimizu, K., Hayashi. Y.,and Ohki, M. (1994). Cancrs R P . ~54, Inaba, T.. Roherts, W. M., Shapiro, L. H., Jolly. K. W.,Raimondi, S. C., Smith, S. D., and Look. A. T. (1992). Science 257, 531-534. Inaba. T.. Shapiro. L. H., Funabiki, ?I, Sinclair, A. E., Jones, B. G.,Ashmun, R. A., and Look, A. T. (1994). Mol. Cell. Biol. 14, 3403-3413. Izraeli, S.. Kovar. H., Gadner, H.. and Lion, T. (1992). Blood 80, 1413-1417. Jeon, 1.. Davis, J. N . , Braun, B. S., Sublett,]. E.. Roussel, M. F., Denny, C:. T., and Shapiro, D. N. (1995).Catirer Rrs. 10, 1229-1234. Kakizuka. A.. Miller, W. H., Jr.. Umesono. K., Warrell, R. P., Jr., Frankel. S. R., Murty, V. V.. Dmitrovsky, E., and Evans, R. M. (3991). Cell 66, 663-674. Kamps. M. P., and Baltimore, D. (1993).Mid. CrfI. Biol. 13, 351-357. Kamps. M. P., Murre, C., Sun, X. H., and Baltimore, D. (1990).Crll 60, 545-555. Kelliher, M., Kriott. A., McLaughlin, J.. Witte, 0. N., and Rosenberg, N. (1991). Mof. Cell. B i d . 11, 4710-4716. Kenned?. M. A., C;orrzalez-Sarmientc,, R.. Kees, U. R., Lampert, F., Dear, N., Boehni, T., and Rabbitts. T. H. (199lj. Pror. iVaf1. Acad. Sci. USA 88, 8900-8904. Koken. M . H., Puvion-Dutilleul. F., Guillemin, M. C., Viron, A., Linares-Cruz, G., Stuurman, N..d e Jong, L., Szostecki, C . , Calvo. F., Chomienne, C., e t a / . (1994). E M B O J . 13, 1073-1083. Kreider, B. L.. Orkin, S. H., and Ihle. J. N. (1993).Pror. Null. Acacl. Sci. USA 90, 64546458. Ladanvi. M., and Gerald, W. (1994). Cancer RPS.54, 2837-2840. Langdon, W. Y.,Harris, A. W., Cory, S., and Adams, J. M. (1986). CeIf 47, 11-18. Larsson. L. (;.< Pettersson, M., Oberg. F., Milsson. K., and Luscher, B. (1994). Oncogene 9, 1247- 1252. LeBrun. L). L. ;inti Clear!. M. L. (1994). Oncogew 9, 1641-1647. Levine. M. S.. and Harding, K. W. (1989). "Genes and Embryos," pp. 39-94, IRL, New York. ., tiajra, A , Claxton, D. F.. Marlton, P., Freedman. M., Siciliano, M. J., and ( : o h i s . F. S. (1993). SciPnre 261, 1041. Lombardi. L.. Newcornb, E. W., and Dalla-Favera. R. (1987). Cell 49, 161-170. Longo, L., Pdndolfi. P. P., Biondi, A.. Rambaldi, A., Mencarelli, A., Lo Coco, F., Diverio, D., Pegoraro, L.. Avanzi, G . , Tabilio, A. el a/.(199O).J. Exp. Med. 172, 1571-1575. Lu, M., Gong, Z. Y.,Shen, W.F., a n d Ho, A. D. (1991). E M B O J . 10, 2905-2910. Lu, Q., Wright, D. D., and Karnps. M. P. (1994). M o l . Cell. Biol. 14, 3938-3948. May, M:. A.. (hhizky. M. L., Lessnick. S. L., Lunsford, L. B.. Lewis, B. C., Delattre, O., Zucnian. J., Thomas, G., and Denny, C. T. (1993a). Proc. Natl. Acad. Sci. USA 90, 5752.5756. Ma!; W. A., Lessnick. S. I.., Braurr. B. S.. Klenisz, hi., Lewis, B. C., Lunsford, L. B., Hronias, R., and Derin:, (:. T. (1993b). ,\fo/. Cell. Bzol. 13, 7393--7398. hlazo, A. M., Huang. D. H.. Mozer. B. A.. and Dawid, I. B. (1990). Proc. Natl. Acad. Sci. US.4 a7,2112-2116. hlrChnis, W.. and Krumlauf, R. (1992). Cell 68, 283-302. Mc(;uirc. E. A , , Hockett. K. D.. Pollock, K. M., Bartholdi, M. F., O'Brien, S. J., and Korsmever, S. J. (1989). Mol. Cell. Biol. 9, 2124-2132. Xlc(;uire, E. A . , Rintoul, C. E.. Sclar. G.M., and Korsnieyer, S. J. (1992).Mol. Cell. Biol. 12, 41X6-419ti.

A DEVELOPMENTAL MODEL OF “MASTER” TRANSCRIPTION FACTORS

55

McKeithan, T. W., Shirna, E. A,, Le Beau, M. M., Minowada, J., Rowley, J. D., and Diaz, M. 0. (1986). Proc. Natl. Acad. Sci. USA 83, 6636-6640. Mellentin, J. D., Smith, S. D., and Cleary, M. L. (1989). Cell 58, 77-83. Meyers, S., Downing, J. R., and Hiebert, S. W. (1993). Mol. Cell. Biol. 13, 6336. Mirro, J., Kitchingrnan, G., Williams, D., et al. (1986). Blood 67, 689. Mitani, K., Ogawa, S., Tanaka, T., Miyoshi, H., Kurokawa, M., Mano, H., Yazaki, Y., Ohki, M., and Hirai, H. (1994). E M B O J . 13, 504-510. Mitelman, F. (1991). “Catalog of Chromosome Aberrations in Cancer.” Wiley-Liss, Inc., New York. Miyoshi, H., Shimizu, K., Kozu, T., Maseki, N., Kaneko, Y., and Ohki, M. (1991).Proc. Natl. Acad. SCZ.USA 88, 10431-10434. Monica, K., LeBrun, D. P., Dedera, D. A., Brown, R., and Cleary, M. L. (1994).Mol. Cell. Biol. 14, 8304-8314. Morishita, K., Parker, D. S., Mucenski, M. L., Jenkins, N. A., Copeland, N. G., and Ihle, J. N. (1988). Cell 54, 831-840. Morishita, K., Parganas, E., Bartholomew, C., Sacchi, N., Valentine, M. B., Raimondi, S. C., Le Beau, M. M., and Ihle, J. N. (1990). Oncogene Res. 5, 221-231. Morishita, K., Parganas, E., Willman, C. L., Whittaker, M. H., Drabkin, H., Oval, J., Taetle, R., and Ihle, J. N. (1992). Proc. Natl. Acad. Scz. USA 89, 3937-3941. Morrissey, J., Tkachuk, D. C., Milatovich, A., Francke, U., Link, M., and Cleary, M. L. (1993). Blood 81, 1124-1131. Mucenski, M. L., McLain, K., Kier, A. B., Swerdlow, S. H., Schreiner, C. M., Miller, T. A., Pietryga, D. W., Scott, W. J., Jr., and Potter, S. S. (1991). Cell 65, 677-689. Mueller, C. R., Maire, P., and Schibler, U. (1990). Cell 61, 279-291. Nagasaka, M., Maeda, S., Maeda, H., et al. (1983). Blood 61, 1174. Nakamura, T., Alder, H., Gu, Y., Prasad, R., Canaani, O., Kamada, N., Gale, R. P., Lange, B., Crist, W. M., and Nowell, P. C. (1993). Proc. Natl. Acad. Scz. USA 90, 4631. Nourse, J., Mellentin, J. D., Galili, N., Wilkinson, J., Stanbridge, E., Smith, S. D., and Cleary, M. L. (1990). Cell 60, 535-545. Nuchprayoon, I., Meyers, S., Scott, L. M., Suzow, J., Hiebert, S., and Friedman, A. D. (1994). Mol. Cell Biol. 14, 5558-5568. Nucifora, G., Begy, C. R., Erickson, P., Drabkin, H. A., and Rowley, J. D. (1993). Proc. Natl. Acad. Sci. USA 90,7784-7788. Numata, S., Kato, K., and Horibe, K. (1993). Leukemia 7, 1441. Nusslein-Volhard, C., and Wieschaus, E. (1980). Nature 287, 795-801. Nusslein-Volhard, C., Frohnhofer, H. G., and Lehmann, R. (1987). Science 238, 16751681. Ogawa, E., Inuzuka, M., Maruyama, M., Satake, M., Naito-Fujirnoto, M., Ito, Y., and Shigesada, K. (1993a). Virology 194, 314. Ogawa, E., Maruyarna, M., Kagoshima, H., Inuzuka, M., Lu, J., Satake, M., Shigesada, K., and Ito, Y. (1993b). Proc. Natl. Acad. Sci. USA 90, 6859. Papavassiliou, A. G. (1995). New Eng1.J. Med. 332, 45-47. Perez-Alvarado, G. C., et al. (1994). Nature Struct. Btol. 1, 388-398. Perkins, A. S., Fishel, R., Jenkins, N. A., and Copeland, N. G. (1991). Mol. Cell. Biol. 11, 2665-2674. Pevny, L., Simon, M. C., Robertson, E., Klein, W. H., Tsai, S. F., DAgati, V., Orkin, S. H., and Costantini, F. (1991). Nature 349, 257-260. Prendergast, G. C., Lawe, D., and Ziff, E. B. (1991). Cell 65, 395-407. Privitera, E., Kamps, M. P., Hayashi, Y., Inaba, T., Shapiro, L. H., Raimondi, S. C., Behm, F.,

56

A . THOMAS LOOK

Hendershot, L., Carroll, A. J., Baltimore, D., and Look, A. T. (1992). Blood 79, 17811788. Pui, C.-H., Behm, F. G., Raimondi, S. C., Dodge, R. K., <;eorge, S. L., Rivera, G . K., Mirro, J., Kalwinsky, D. K., Dahl, G. V., Murphy, S. B., Crist, W. M., and Williams, D. L. (1989). New Eiig1.J. M r d . 321, 136-142. Rabbitts, T. H. (1991). Cell 67, 641-644. Rabbitts, T. H. (1994). ,Voture 374, 143-149. Rabbitts, T, H., and Boehm, T. ( I Y Y 1). Adv. Itiiiumol. 50, 119-146. Rabbitts, T. H., Forster, A., Larson, R., and Nathan, P. (1993). ,Vature Genet. 4, 175-180. Raimondi, S. C. (1993). Blood 70, 2237-22.51. Raju, R. E. (1993). hfech. Dev. 44, 51-64, Rappold, G. A . . Hameister, H., Cremer, T., P I al. (1984). E M B O J . 3, 2951-2955. Rauskolb, C.. Peifer, M., and Wieschaus, E. (1993). Cell 74, 1101-1 112. Rivera, G. K., Raimondi, S. C.. Hancork, M. L., Behm, F. G., Pui, C., Abromowitch, M., Mirro, J.. Jr.> Ochs, J. S., Look, A. T., Williams, D. L., Murphy, S. B., Dahl, G . V., Kalwinsky, D. K . , Evans, W. E., Kun, L. E.. Simone,,J.V., and Crist, W. M. (1991). Lancet 337,6 1-66. Roberts, C. W. M . , Shutter, J. R . , and Korsnieyer, S. J. (1994). Nature 368, 747-749. Rowle), J. D., Aster, J. C., and Sklar, J. (l993)./. Am. M e d . Assoc. 270, 2331-2337. Royer-Pokora, B., Loos, U., and Ludwig, W.D. (1991). Oncogene 6, 1887-1893. Schichman. S. .4., Caligiuri. M. A., Gu, Y., Strout, M . P., Canaani, E., Bloomfield, C . D., and Croce, C. M .(1994a). Proc. h'atl. Arad. Sci. C'SA 91, 6236-6239. Schichman, S. A., Caligiuri, M .A., Strout, hi. P., Carter, S. L., Gu, Y., Canaani, E., Bloomfield, C. D., and Croce, C. M. (1994b). Cnnrn- Res. 54, 4277-4280. Scott, E. W., Simon, M . C., .4nastasi, J., and Singh, H . (1994). Sczenre 265, 1573-1577. Shapiro, D. N., Sublett, J. E.. Li, B., Downing.]. R., and Naeve, C;. U'. (1993). CancerRes. 53, 5108-51 12. Shima, E. A, Le Beau, M. M.,McKeithan, T. W., Minowada, J,, Showe, L. C., Mak, T. W., Minden. M. D., Rowley. J. D., and Diaz, M. 0. (1986). Proc. Natl. Acad. Sci. USA 83, 3439-3443. Shimizu, K., Miyoshi, H., Kozu, T., Nagata, J., Enomoto, K., Maseki, N., Kaneko, Y., and Ohki. M. (1992). Caiirer Re,. 52, 6 9 4 5 6 9 4 8 . Sorensen, P. H . . Lessnick. S. L., Lopez-Terrada. D., Liu. X. F., Triche, T. J., and Denny, C:. T. ( 1Bt)4a). Suture Grnct. 6, 146- 15 I . Sorensen. P. H., Lessnick, S. L., Lopez-Terrada, D., Liu, X. F., Triche, T.J., and Denny, C. T. (1994b). AVatiweG P ~ P6, ~146-151. . Stong, R. C., Korsnieyer, S. J.. Parkin. J. L., Arthur, D. C . , and Kersey, J. H . (1985). Blood 65, 21. ?Iaub, R . , Kirsch. I . , Morton, C . , Lenoir, G., Swan, D., Tronick. S.,Aaronson, S., and Ledel-, P. (1987). P w r . .Vat/. ilcnd. Sri. CSA 79, 7837-784 1. Taub, K., Kelly, K.. Battey, J., Latt, S., Lenoir, G. hi., Tantravahi, U., Tu, Z., and Leder, P. ( 1984). Cell 37, 5 1 1-520. Tkachuk. D. C . . Kohler, S., and Clear?, M. L. (1992).Crll71, 691-700. T'sai, F. Y.,Keller. G., Kuo. F. C.. Weiss. X i . , Chen, J., Rosenblatt, M.,Alt, F. W., and Orkin, S. H. (1994). , V ~ ~ U371, Y P 221-226. Valge-Archer. V. E., Osada, H.. Warren, r\. J., Forster, A,, Li, J., Baer, K.,and Kabbilts, T. €-I. (1994). Proc. Satl. Acnd. Sri. L'S.4 91, 8617-8621. Van Dijk, M. A,. and Murre. C. (1994). Cell 78, 617-624. Van Dijk, M. .4., \'ool-hoeve. P. X i . , and Murre, C . (1993). P+oc. ,Vat(. Acad. Sc7. USA 90, 6061-(506.3.

A DEVELOPMENTAL MODEL OF "MASTER" TRANSCRIPTION FACTORS

57

Van Etten, R. A,, Jackson, P., and Baltimore, D. (1989). Cell 58, 669-678. W a t h a n , I . , Li, J., Bash, R. O., Forster, A., Osada, H., Rabbitts, T. H., and Baer, R. (1994). EMBO J . 13,4831-4839. Waing, S., Wang, Q., Crute, B. E., Melnikova, I. N., Keller, S. R., and Speck, N. A. (1993). EMRO J. 13, 3324. Warrell, R., Jr., Frankel, S., Miller, W., Jr., Scheinberg, D., Itri, L., Hittelman, W., Vyas, R., Andreeff, M., Tafuri, A., Jakubowski, A., GabriloveJanice, Gordon, M., and Dimitrovsky, E. (1991). New Engl. J . Med. 324, 1385-1393. Warrell, R. P., Jr., Maslak, P., Eardley, A., Heller, G., Miller, W. H., Jr., and Frankel, S. R. (1994).Leukemia 8, 929-933. Warren, A. J., Colledge, W. H., Carlton, M. B. L., Evans, M. J., Smith, A. J. H., and Rabbitts, T. H. (1994). Cell 78, 45-57. Weintraub, H. (1993). Cell 75, 1241-1244. Weis, K., Rambaud, S., Lavau, C., Jansen, J., Carvalho, T., Carmo-Fonseca, M., Lamond, A., and Dejean, A. (1994). Cell 76, 345-356. Xia, Y., Brown, L., Yang, C. Y., Tsan, J. T., Siciliano, M. J., Espinosa, R., 111, Le Beau, M. M., and Baer, R. J. (1991).Proc. Natl. Acad. Scz. USA 88, 11416-11420. Yoshihara, T., Inaba, T., Shapiro, L. H., Kato, J., and Look, A. T. (1995).Mol. Cell. B i d . 15, 3247-3255. Zervos, A. S., Gyuris, J., and Brent, R. (1993). Cell 72, 223-232. Ziemin-van d e r Poel, S., McCabe, N. R., Gill, H. J., Espinosa, R., Patel, Y., Harden, A., Rubinelli, P., Smith, S. D., Lebeau, M. M., Rowley, J. D., and Diaz, M. 0. (1991).Proc. Natl. Acad. Sci. USA 86, 10735. Zinszer, H., Albalat, R.,and Ron, D. ( 1 994). Genes Dcv. 8, 25 13-2526. Zucman, J., Delattre, O., Desmaze, C., Epstein, A. L., Stenman, G., Speleman, F., Fletchers, C. D., Aurias, A., and Thomas, G. ( 1993a). Nature Genet. 4, 34 1-345. Zucman, J., Melot, T., Desmaze, C., Ghysdael, J., Plougastel, B., Peter, M., Zucker, J. M., Triche, T. J., Sheer, D., Turc-Carel, C., et al. (1993b). EMBO J . 12, 4481-4487.

This Page Intentionally Left Blank

PATHWAYS OF CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS Bernard Dutrillaux URA 620, CNRS-lnstitut Curie, 75231 Paris, France

I. Introduction 11. Colorectal Adenocarcinoma A. Definition of the Different Cytogenetic Tumor Types B. Relationship between Cytogenetic Tumor Type and Other Parameters C. Cytogenetic Tumor Type and Genetic Predisposition D. Pathways of Genomic Alteration and Tumor Type 111. Other Epithelial Tumors A. Breast Cancer B. Non-Small-Cell Lung Cancer C. Endometrial Cancer IV. Microsatellite Instability in Epithelial Tumors A. Breast Cancer B. Lung Cancer C. Endometrial Cancer V. Concluding Remarks References

1. Introduction

Chromosome changes in cancer cells attracted much attention when it was shown that the alteration of genes, oncogenes, or tumor suppressors could be directly or indirectly related to a given chromosome rearrangement [see Sandberg (1990) for historical background]. A major goal was then to search for new cancer genes, whose existence was suggested by the observation of a recurrent chromosome rearrangement in a malignancy. This research was successful and led to the discovery of many such genes. However, besides hematological malignancies, which can be regarded as model systems since a specific chromosome alteration frequently characterizes a pathology, solid tumors, and epithelial cancers in particular, rather are characterized by a large number of chromosome alterations, none of which was shown to be specific. T h e concept that some deletions could unmask recessive mutations aLffecting tumor suppressor genes was progressively admitted by most investigators, even if it remains rarely fully demonstrated. In retinoblas59 ADVANCES IN CANCER RESEARCH. VOL. 67

Copyright 0 1995 by Academic Press, Inc. All rights of reproduction in any form reserved

60

BERNARD DUTRILLAUX

toma, where RB1 gene from chromosome 13 is the cause, the loss of chromosome 13 is much less frequent than in some cancers, such as breast or lung adenocarcinomas. In retinoblastoma, the loss of RB 1 function parallels allelic but less strictly chromosomal losses (Cavenee et al., 1983; Chaum et nl., 1984),whereas in breast cancer, chromosome 13 loss (Dutrillaux et nl., 1990) may coexist with RB1 activity, which may be explained by the role of another susceptibility gene, such as BRUSH-1 on chromosome 13 (Schott et al., 1994). Oversimplifications must be rejected for interpreting the meaning of the TP53 gene alterations. The 17p arm, which carries the gene, is frequently deleted from epithelial tumors, in which TP53 gene is most frequently altered by mutation (Baker et al., 1989). In colorectal carcinoma, for instance, in spite of a high recurrence of both events, the 17p arm may be deleted in the absence of TP53 mutation (Muleris, Remvikos, and Dutrillaux, unpublished observations). This suggests that the deletion can be selected for a reason other than the loss of p53 function. In epithelial tumors, chromosome alterations are numerous and d o not occur at random, but none is specific for a given tumor type. Some tumor types, such as breast cancer, undergo a very high rate of chromosome rearrangements, while others, such as endometrial carcinoma, d o not. T h e same is true for the occurrence of endoreduplications, chromosome malsegregation, and gene amplification. One aim of this report is to show that the occurrence of these alterations is coordinated, suggesting that the first alterations determine the nature of the following. A major question is the identification of the selective pressure leading the tumor genome to undergo chromosome losses rather than gains or vice versa. This will not be developed here, but we shall just recall our interpretation, which has not been confirmed by other groups: the selective pressure would be determined by early metabolic changes, which precede chromosome alterations (Bravard et al., 1992: Hoffschir et al., 1993). T h e existence of distinct pathways of chromosomal evolution in human solid tumors was first described in colorectal adenocarcinomas (Muleris PI al., 1988). It was proposed that a subset of tumors evolved by structural rearrangements of chromosomes, leading to imbalances, in particular deletions of chromosome arms and that another subset of tumors progressively increased their chromosome numbers by somatic malsegregations, with a lower occurrence of structural rearrangements. ‘The first group of tumors was called the “monosomic” type and the second the “trisotnic” type. A third group is composed of tumors with normal karyotype. At the same time, Japanese investigators described two distinct karyo-

CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS

61

typic patterns in neuroblastomas (Kaneko et al., 1987; Hayashi et al., 1989), which were associated with distinct prognoses. These authors placed no emphasis on the pathways leading to the different karyotypes observed, but the reassessment of their data indicates that neuroblastomas of good prognosis follow a “trisomic” evolution and those of adverse prognosis a “monosomic” .evolution. This suggests that these purely cytogenetic alterations parallel a number of other genetic alterations of tumors cells and also pathological characters. T h e aim of this report is to propose that each epithelial tumor follows one of these routes and that other biological parameters may be directly related to it. T h e example of colorectal cancer will be developed first, and data from various other tumors will be reassessed in relation to this classification. Emphasis will be placed on the possible relationship between tumor evolution and microsatellite instability.

II. Colorectal Adenocarcinoma There are numerous publications about the cytogenetics of colorectal cancer, with large variations in the types of chromosome alterations described (Reichmann et al., 1981; Couturier-Turpin et al., 1982; Ochi et al., 1983; Muleris et al., 1985, 1990b; Levin and Reichmann, 1986; Bardi c t al., 1993). Indeed, if our hypothesis that subsets of tumors with distinct chromosomal patterns exist is to be proved, it is necessary to emphasize their respective characteristics, which may be hidden by mixing all the data. Nevertheless, to compare cytogenetic and molecular data, it is useful to have a short statistical description of all cases. I n a series of 100 cases (Muleris et al., 1990b), the most frequently observed aberrations were losses of the 17p arm and chromosome 18, in about 75% of cases each. The other alterations are indicated in Table I. When compared to molecular data on allelic losses (LOH), very similar frequencies of LOH were reported for both the 1’7p and 18q arms (Delattre et al., 1989). This suggests that these alterations are representative of this type of cancer. When the distribution of these two alterations is considered in the 100 cases, they are too frequently observed together: in 70 instead of (- 17p)76 x (- 18)78/100 = 59 expected cases if the two anomalies were independent. The same conclusion about concerted losses was obtained for LOH (Delattre et al., 1989). When the other deficiencies are considered (Table I), they are all much more frequent when the 17p arm and chromosome 18 are deficient together than when not. In contrast, when chromosome gains are considered, all are much less

62

BERNARD DUTRILLAUX

TABLE I PERCENTAGES OF THE

MOST FREQUENT IMBALANCES

O B S E R V E D IN A SERIES OF

Chromosome irn balances

- 18 - 17p

del( 1Kp3) - 8P del( 1 -Y del(.i)(q2) - 14 - 4q -21 - 1.5

-X(late) - 22 - 9q -11 +>((males) + 13 + 20 +>((earl)) +8q +7

+ 12

.411 t)

100 C o L O R E C r A L

CANCERSn

Monosomic ‘Y Pe

Trisornic type

78 76

100 100

32 27

33

51

69 66

18 18

49

65

49 47 17

65 60 65

14 37 18 0

44

61 60 .i 9 . 7i

pes

44 42 41

35 33 28 . 9i . i 1 43 43 40

3 17

48 45 38 74 47 39 50 38 19 4

4 9 0

25 4 9 4 62 77 68 62 54 86 64

n I n tumors of the female. early and law replicaiing X’s were identified by BrdU incorporation [from Muleris et al. (1990b)l.

frequent when the 17p arm and chromosome 18 are deficient together than when not, except for the gain of early replicating X, which was about as frequent in the two groups of tumors. A. DEFINITION OF THE DIFFERENT CYTOCENETIC

TUMOR TYPES On the basis of chromosome changes, it was proposed that two distinct pathways of evolution existed in colorectal cancers, which can be defined as follows (Muleris et al., 1988, 1990b).

1. A large subset of tumors, about 70% of cases, undergoes structural rearrangement of a number of chromosomes. The chromosomes most frequently involved are numbers 17, 1, 8, 10, 5 , 4, and 9, in decreasing order of involvement. These rearrangements lead to deletions of the

CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS

63

following arms, 17p, lp, 8p, 10 (intercalary), 5q (intercalary), 4q, and 9q, and less frequently to duplications of the 8q and 17q arms. Mitotic malsegregations also occur, leading to whole losses of chromosomes 18, 14, 15,21, and 22 and of sex chromosomes (late replicating X and Y). Some gains may also involve, in decreasing order, chromosomes X (early replicating), 13, 20, and 7. On the whole, these alterations lead to a decrease in the total number of chromosomes, most rearrangements being whole arm translocations with the loss of one of the two derivative chromosomes. T h e lowest karyotypic formula observed was 4 1. These tumors, with a hypodiploid karyotype, have, however, a strong tendency to undergo endoreduplicat ions, leading to the formation of hypotetraploid subclones. These subclones have variable proliferative potentials, but are sometimes better than the hypodiploid cells from which they originated. Thus, in about 40% of cases, the coexistence of hypodiploid and hypotetraploid clones is observed, and in 60% of cases, only the hypotetraploid clone is conserved. It was shown that the same types of allelic losses were observed in hypodiploid and hypotetraploid tumors. This, as also suggested by the presence of duplicated derivative chromosomes, demonstrates that hypotetraploid tumors derive from hypodiploid tumors (Muleris et al., 1990a). T h e range of chromosome numbers is quite large for tumors having undergone endoreduplication, from about 60 to more than 120. This is related to the frequent whole chromosome losses occurring after endoreduplication, which is evidenced by the multiple cell to cell variations observed within a given tumor. Thus, the tendency for chromosome loss is much stronger after than before endoreduplication, and chromosome rearrangements prevail. This leads to a decrease in the formula from hypotetraploidy to pseudotriploidy without accumulation of structural rearrangements. A small number of tumors undergo a second endoreduplication, which increases their formulae to up to 120 chromosomes. T h e data obtained on chromosome numbers are well correlated with DNA content measured by flow cytometry (Remvikos et al., 1988b). Due to the basic mechanisms leading to all of these karyotypic variations, especially chromosome losses and deletions, these tumors were called monosomic, independent of their total number of chromosomes. 2. A subset of 20-25% of tumors has karyotypic alterations of another nature. Few structural rearrangements and chromosome losses occur. In contrast, chromosome duplications are frequent, leading to a progressive increase in the karyotypic formulae, which range from 47 to 58 (Muleris et al., 1990b). The most recurrent alterations were, in decreasing order of frequency, +7 (86%), +13, +20, +12, + X (early replicat-

64

BERNARD DUTRILLAUX

ing), and +8q (54%) (Table I). In these tumors, polyploid side lines are exceptional, indicating that endoreduplication is a rare event. This interpretation is strengthened by the observation that all near-tetraploid tumors carry rearrangements and deletions characterizing the monosomic type described earlier. These tumors were called trisomic because trisomies of various chromosomes are their most frequent alterations. 3. Finally, in about 5-7% of tumors, only normal karyotypes are observed. Such a finding must be considered cautiously, since it may simply result from obtaining karyotypes from noncancerous cells. However, in the cases we studied, flow cytometry showed that DNA indices were always 1, whereas in tumors with aneuploid karyotypes, DNA indices varied in relation to the aneuploidy (Remvikos et al., 198813). As we shall see, these tumors with normal karyotypes share several characteristics with trisomic type tumors. Their malignant character was demonstrated by their capacity for xenografting in "nude" mice.

B. RELATIONSHIP BETWEEN CYTOCENETIC TUMOR T Y P E A N D OTHER PARAMETERS This distinction of tumor subsets, based on purely cytogenetic data, may serve as a guideline for studying other parameters. 1 . Tumor Localzzataon

In the sample of 100 tumors published by Muleris c7t al. (1990b), 14 had a proximal, 42 a distal, and 44 a rectal localization. Almost all tumors of the proximal colon either were of the trisomic type or had a normal karyotype, and almost all tumors of the distal colon were of the monosomic type. In the rectum, the various cytogenetic types were observed in the expected proportions. This observation fits with earlier findings of Reichmann et al. (1981), who described three tumors with normal karyotypes in the proximal colon and with data on LOH for chromosomes 5, 17p, and 18 (Delattre et al., 1989). 2. Metabolism of Thymzdine

Early studies pointed out that unusual activities of thymidylate synrhase (TYMS) and thymidine kinase ( T K ) existed in some colorectal cancers (Weber et al., 1978). The activity of these two enzymes was studied in relation to the chromosomal pattern of colorectal cancers. TYMS is mapped on chromosome 18, which is always deficient in monosomic tumors, and T K is mapped on the 17q arm, which is very rarely deficient but frequently duplicated in monosomic and occasionally in trisomic tumors, suggesting a possible imbalance of enzyme activity related to

CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS

65

gene dosage (Dutrillaux and Muleris, 1986). By using xenografted tumors devoid of stromal cells, which preserved characteristic chromosome patterns (Lefranqois et al., 1989), it was shown that TYMS activity was low o r very low and T K activity high in the monosomic type tumors, whereas TYMS activity was much higher and T K activity lower in the trisomic type tumors (Bardot et al., 1991). This suggested that a gene dosage effect might be at least partly responsible for these unusual metabolic patterns, which was confirmed by mRNA analyses, both in xenografts and in fresh tumors, by an in situ approach (Lasserre et al., :1994a,b). Other enzyme activities were also shown to depend on chromosome alterations, such as UMPK (uridine monophosphate kinase, mapped on l p arm) (Bravard et al., 1991). Thus, the metabolic pathway of thymidine 5-phosphate synthesis is strongly correlated with the cytogenetic type of colorectal cancer, the trisomic type of tumors and tumors with normal karyotypes being more alike than the monosomic type.

(2. CYTOGENETIC TUMOR TYPEAND GENETIC PREDISPOSITION Two categories of genetic predisposition exist, according to the presence of multiple polyps [familial adenomatous polyps (FAP)] or their absence [hereditary nonpolyposis coli cancers (HNPCC)]. 1 . Tumors in FAP Patients

T h e heterozygote constitutional involvement of the APC gene mapped on the 5q arm is well documented in FAP patients (Bodmer et al., 1987; Groden et al., 1991). In cancer cells from these heterozygote patients, the normal APC allele, present in other tissues, is either lost or mutated, suggesting a recessive determinism for tumor development (Nagase and Nakamura, 1993). Most losses seem to result from the deletion of the 5q arm, which is cytogenetically observed in about 50% of cases, independently from FAP predisposition (Muleris et al., 1990b; Table I). The frequency of this deletion in the monosomic type (about 0.60) is threefold higher than in the trisomic type (about 0.18). Indeed, it does not occur in tumors with normal karyotypes. In the short series of six tumors from FAP patients we studied (a seventh case was ambiguous), all were of the monosomic type (Muleris and Dutrillaux, unpublished observations). Thus, most cancers developed in FAP patients are of the monosomic type and have a deletion of 5q arm.

66

BERNARD DbTRILLAUX

2. Tumors in HNPCC Patients

Hereditary nonpolyposis colorectal cancer patients exhibit an excess of tumors, principally colorectal adenocarcinomas. Their overabundance in the proximal bowel is questionable. In HNPCC families, there is also an increased occurrence of endometrial cancers (Lynch et al., 1993). Several HNPCC genes have now been identified by positional cloning. They correspond to mismatch repair genes: hPMSl, hPMS2, hMSH2, and hMLH 1 localized on chromosomes 2, 7, 2, and 3, respectively (Aaltonen et al., 1993, 1994; Bronner et al., 1994; Nicolaides et al., 1994). There is still no real consensus about the definition of HNPCC patients. Ideally, they could be defined by the presence of a mutation of one of the aforementioned genes, but the possibility that other as yet unknown genes are involved prevents such definition. Whether all HNPCC patients carry a mismatch repair gene mutation, and whether all carriers of such a mutation can be regarded as HNPCC patients remain open questions. I t has been established that these mutations induce multiple replication errors (RER), leading to microsatellite instability: RER+ phenotype. It is not yet clear whether or not the mutation always passes to homozygosity in tumor cells, although it was found (Papadopoulos et al., 1994). The frequency of HNPCC patients is not accurately known: its estimate ranges from 0.01 to 0.1 among patients affected by a colorectal cancer, depending on the definition retained by the authors. Given the high occurrence of colorectal cancers, HNPCC would be one of the most frequent genetically determined diseases. At present, a majority of the tumors from identified HNPCC patients were RER+, but it seems that the frequency of RER+ tumors largely exceeds that of HNPCC patients: it is observed in about 20% of all colorectal cancers. Considering that tumors from HNPCC patients frequently had normal karyotypes (Muleris et al., 1995) and that the RER+ phenotype was associated with DNA-diploid tumors without allelic losses (Thibodeau et al., 1993), we looked for a correpondence between karyotype and RER phenotype. In the short series studied (13 cases of the monosomic type and 10 cases of the trisomic or normal type), microsatellite instability was observed in tumors of either the trisomic type or with a normal karyotype, but not in those of the monosomic type (Remvikos et al., 1995a,b). Thus, most tumors from APC patients belonging to the monosomic type and most of those from HNPCC patients belonging to the normal or trisomic type follow quite different pathways as regards chromosome alterations. These pathways are similar to those followed by other colorectal cancers, occurring in the majority of patients who are not known to

CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS

67

be genetically predisposed. Our hypothesis is thus that two major pathways exist in the oncogenic process of colorectal cancers. Di. PATHWAYS OF GENOMIC ALTERATION AND TUMOR TYPE

T h e monosomic type of tumors involves many structural rearrangements of chromosomes, leading to the deletion of a number of chromosome arms. Some of these deletions are assumed to unmask occasional recessive mutations, more or less specific for a given tumor type. In colorectal cancer, those of the 5q, 17p, 18p, and 18q arms are already known to be associated with mutation or alteration of the expression of genes such as APC, TP53, TYMS, and DCC (Fearon et al., 1990), respectively, which may be regarded as tumor suppressor genes. Other tumor suppressor genes are likely to exist. For instance, such genes are likely to be located on the short arms of chromosomes 1 and 8, which are very frequently deleted. Besides their chromosome “instability,”these tumors are of the RER- phenotype: microsatellite and chromosome instability thus appear to belong to two different mechanisms. T h e causes of the chromosomal instability are not known, neither whether this instability corresponds to a real increase in the occurrence of structural rearrangements nor whether occasional rearrangements are simply efficiently retained by a strong selective pressure when they induce imbalances favorable to tumor growth. The involvement of hypomethylation as a factor of instability was demonstrated by in vitro experiments (Almeida et al., 1993; Kokalj-Vokac et al., 1993), but not in tumor cells. Strong imbalances in the nucleotide pools, as suggested by the unusually low ratio of thymidylate synthase to thymidine kinase activities (Bardot et al., 1991), may also contribute to genome instability in monosomic type tumors. Telomere shortening, demonstrated in senescent and immortalized cells, is also a major cause of chromosome instability (Broccoli and Cooke, 1993) by the formation of dicentrics, which further undergo breakage-fusion between the two centromeres. This may lead to derivative chromosomes composed of the two arms not involved in the fusion. It may also lead to duplication deficiencies, resulting in the amplification of one sequence and the deletion of others. Data are, however, too limited to correlate telomere shortening and karyotype allterations in human tumors (Schmitt et al., 1994). Indeed, alterations of the TP53 gene, which is very frequent in colorectal cancer and associated with the deletion of the 17p arm in the monosomic type tumors, may also be a cause of genomic instability (Hartwell, 1992; Yin et al.,

68

BERNARD DUTRILLAUX

1992). However, the chronology of the occurrence of 17p arm deletions and TP53 mutations is not firmly established. Our unpublished data (Muleris, Remvikos, and Dutrillaux) suggest that some 17p arm deletions can occur in the absence of TP53 mutation, which prevents any generalization about the role of this gene in chromosome instability. T h e tumors with a normal karyotype indeed have a completely different evolution. They are observed in about 7% of cases, a rate quite similar to that expected for HNPCC patients. Since these tumors have an RER+ phenotype (Remvikos et al., 1995b), they may largely correspond to tumors from HNPCC patients. This is in agreement with clinical data since, in the short series we studied, such tumors were frequently observed in young patients or in familial cases. In cases with a germinal mutation, the alteration of the remaining normal allele would be sufficient to induce the tumoral process. Since no chromosome alterations occur during the tumor progression, it may be assumed that the RER+ phenotype, due to the suppression of one of the mismatch repair genes, induces many mutations, including one o r a few dominant ones, which is sufficient for enabling a complete tumorigenic process. T h e trisomic type tumors share many characteristics with tumors with a normal karyotype, such as preferential localization in the right colon and similar metabolic patterns. Indeed, they are also of the RER+ phenotype, which suggests that their genome also carries mutations of one of the mismatch repair genes. During tumor progression, they do not accumulate many structural rearrangements of chromosomes, but increase their chromosome numbers by progressive gain of apparently normal chromosomes. The lack of deletion suggests that their genomic instability, as demonstrated by their RER+ phenotype, leads to nonrecessive mutations. The mechanisms leading to their chromosome gains are as yet unexplained, although the metabolic pressure and gene dosage effect are assumed to play a role (Lasserre et al., 1994a,b).

Ill. Other Epithelial Tumors A. BREASTCANCER A large heterogeneity exists in the literature regarding the karyotypes of breast cancer cells. Obviously, the results largely depend on culture conditions. T h e harvesting of first division metaphases (less than 4 days and control of divisions by BrdU incorporation) exhibits monoclonal and highly rearranged karyotypes, whereas more prolonged cultures

CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS

69

andlor addition of growth factors provide a variety of karyotypes that are frequently polyclonal, with few or no chromosome rearrangements (F’andis et al., 1993, 1994). It is not our purpose to discuss these differences, and we shall limit our analysis to data obtained from first division metaphases in vitro, for which we studied more than 200 cases under the same conditions used for colorectal cancers. A statistical analysis of 113 cases with abnormal karyotypes was reported, and chromosome data were compared to various other biological parameters such as DNA content (Dutrillaux et al., 1991) and proliferative (Remvikos et al., 1992) and hormonal (MagdelCnat et al., 1992, 1994) status. To provide an analysis comparable to other tumors, we have reassessed these 113 tumors and added 7 1 new cases with abnormal karyotypes. T h e 42 tumors for which only normal chromosomes were observed are not considered here, because it could not be demonstrated whether o r not these karyotypes were representative of tumor cells. Data 011 chromosome numbers and numbers of structural rearrangement are provided by two histograms (Fig. 1). Almost all tumors had variable chromosome numbers, but it was possible to define a modal number, which is considered here. These modal numbers ranged from 35 to 128. I n a proportion of cases, subclones with different levels of ploidy existed: we considered those with the lowest ploidy. In all of the cases, tumors were monoclonal, but variations could be attributed to the presence of subclones. As shown in Fig. la, near-diploid (41-50 chromosomes) tumors were frequent (42%). Highly hypodiploid tumors (9%) very frequently had hyperploid subclones, derived by endoreduplication. Hyperploid tumors (49%) had a wide distribution of chromosome numbers, but most had between 61 and 80 (32%).Only 4% were hypertetraploid. The distribution of the numbers of rearranged chromosomes in relation to ploidy is informative (Fig. lb). To estimate these numbers, we counted, in karyotypes with a chromosome number close to the mode, all abnormal chromosomes, but only once when they were in two or more copies. As almost all rearrangements were unbalanced and resulted in one derivative chromosomes, the number of rearranged chromosomes was fairly representative of the number of rearrangements that occurred. The number of rearranged chromosomes increases from near-diploidy to hypodiploidy. We showed that, in these tumors, losses were directly proportional to chromosome rearrangements. This number is higher in near-tetraploid tumors (81-90) and increases when tumors are less hyperploid. Hypertetraploid tumors have the highest number. This evolution, which was characterized in detail (Dutrillaux et al.,

70

BERNARD DUTRILLAUX

a

42

n=184

V

41< 40

/

50

5

60

70

"

80

Chromosome number

b

Chromosome number FIG. 1 . Breast cancer. (a) Distribution (percentages) of modal chromosome numbers in 184 tumors with abnormal karyotypes. In the same study, only normal karyotypes were found in 42 other tumors. (b) Mean numbers of rearranged chromosomes (averages and standard deviations), calculated in metaphases with a modal chromosome number, in the 184 tumors.

1991), recalls that of the monosomic type of colorectal cancer, defined earlier, but with the following differences: 1. The chromosome segments most frequently involved in rearrangements, principally deletions, differ. They are, in breast cancers, the 16q, 17p, lp, 8p, 1 lq, 1 lp, 6q arms. Duplications principally involve lq and 8q arms (Dutrillaux et a/., 1990). 2. T h e rate of rearrangements is much higher in breast cancer, and their occurrence seems to be continuous. 3. T h e tendency to endoreduplication is stronger in colorectal cancer and less correlated to the occurrence of chromosome rearrangements, but double endoreduplications may be more frequent in breast cancer. 4. For equivalent numbers of chromosomes, the DNA ploidy is high-

CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS

71

er in breast than in colorectal cancer. This is due to the presence of more large derivative chromosomes in breast cancer (Remvikos et al., 1988a). 5. Gene amplifications are more frequent in breast than in colorectal cancers. They are characterized by hsr (homogeneously staining regions) but not by dmin (double minutes). hsr occur in more than half of the cases of breast cancers (Saint-Ruf et al., 1991) and involve a large variety of chromosome segments (Kallioniemi et al., 1994; Muleris et al., 1994). They are not induced by therapeutics and may occur in tumors with few chromosome alterations, but are more frequent in tumors with rnany alterations. 6. T h e increase in chromosome alterations is strongly correlated with adverse pronostic factors, such as cell proliferation (Remvikos et al., 1992), loss of hormonal receptors (Magdelenat et al., 1992), histological grading (Dutrillaux et al., 1991), and young age (Remvikos et al., 1995a) in breast cancer. Except for the deletion of the 17p arm, we could not find such correlations for colorectal cancer (Muleris and Dutrillaux, unpublished observations).

I n conclusion, karyotype alterations of breast cancer cells are much more important than those of colorectal cancer cells, but follow the same pathway as the monosomic type of colorectal cancer. Their continuous aind higher occurrence may explain their strong correlation with prognostic factors. 13. NON-SMALL-CELL LUNGCANCER

T h e same type of statistical study as before was performed on lung c:ancer, although less data were available. We considered the series published by Testa et al. (1994), our own data (Viegas-P4quignot et al., 1990; Flury-Herard et al., 1992), and those of Lukeis et al. (1990). This makes a total sample of 83 tumors, principally composed of adenocarcinomas and squamous cell carcinomas. They were analyzed together after their chromosome evolutions were compared and found to be quite similar. T h e distribution of chromosome numbers is different from that of breast cancer (Fig. 2a). A majority of karyotypes (61%) possess between 51 and 70 chromosomes (this percentage was 23% for breast cancers), and only 8% are near-diploid. This difference is probably indicative of a real biological difference, but it must be stressed that, in contrast with breast cancer, these numbers were not the modal ones, which were not given in published data, but the means between extreme values, which was the only information provided about chromosome numbers. T h e distribution of the number of rearrangements (Fig. 2b) was, in

72

BERNARD DUTRILLAUX

35

a

n-83

41< 40

50

60

70

80

90 >90

Chromosome number

b

p 41c 40

!

Chromosome number

FIG. 2. Non-small-cell-lung cancer. (a) Distribution (percentages) of-chromosome numbers in 83 tumors with abnormal karyotypes. (b) Mean numbers of rearranged chromosomes (averages and standard deviations). Data were obtained by calculating the averages hetween extreme values given in karyotypic formulas from literature.

contrast, quite similar to that of breast cancer: their number increases from near- to hypodiploidy and from hypotetraploidy (7 1-80) to hyperdiploidy (51-60). T h e major difference was that the highest rate of rearrangement was observed in near-tetraploid tumors. These similarities and differences with breast cancer can be explained by the relative importance of the occurrence of chromosome losses and rearrangements. Both lung and breast tumors evolve as monosomic type tumors. However, in lung as in colorectal cancer, chromosome losses are less directly related to structural rearrangements than in breast cancer: there are proportionally more losses but fewer rearrangements. Starting from diploid?, the decrease in chromosome number in lung cancer is apparently fast (very few near diploid tumors) and can reach 35 chromosomes, and endoreduplication very frequently occurs on hypodiploid tumors as in breast cancer. This leads to the following stage with less than 80 chromosomes. Both rearrangements and losses continue to accumulate. Chromosome numbers decrease to less than 70. Tumors with 3 - 6 0 chromosomes are derived from hypotetraploid tumors. A second endoreduplication occurs, leading to hypertetraploidy. T h e process of chromosome rearrangements and losses continues, so that hypotetra-

CHROMOSOME ALTERATION I N HUMAN EPITHELIAL CANCERS

73

ploid tumors (81-90 chromosomes) have undergone two endoreduplications, many structural rearrangements, and many losses by mitotic malsegregations. T h e pattern of deletions is similar to that of neither colorectal nor breast cancer. Chromosome 13 is frequently missing, as well as the 9p, :3p, 17p, 8p, and 6q arms. The distribution of deletions may vary according to the histological type. For instance, the 3p deletion was much more frequent in squamous cell carcinomas than in adenocarcinomas (ViegasRquignot et al., 1990; Testa et al., 1994). Gains involve chromosome 7 and the l q arm principally, a characteristic shared with many other solid tumors. There is no firm data about any relationship between chromosome alterations and prognostic factors in lung cancer. In the series of Testa et tzl. (1994), there is apparently no relationship between the histological stage and chromosome alterations. (2. ENDOMETRIAL CANCER Cytogenetic data on endometrial adenocarcinomas are still limited. Fortunately, they are quite homogeneous and simple, which enables a statistical analysis of a small number of cases by adding the short series published since 1985 (Fujita et al., 1985; Couturier et al., 1986, 1988; IDutrillaux and Couturier, 1986; Jenkyn and McCartney, 1986; Yoshida et al., 1986; Gibas and Rubin, 1987; Huber et al., 1990; Tharapel et al., 1991; Shah et al., 1994). Chromosome numbers were almost all (33/38) between 46 and 50; the !j others ranged from 52 to 69 (Fig. 3). Most karyotypes (26/38) had less

Number of rearrangements

00-1

2-3

>3

n=38

46

50

I

55

"

1

1

I 65

I

Chromosome number

FIG. 3. Endometrial cancer. Distribution of chromosome numbers in 38 tumors, in irelation to the numbers of structural rearrangements.

74

BERNARD DUTRILLAUX

than two and only four had more than three structural rearrangements. Most alterations were numerical, and rearrangements generally led to gains (principally of the lq arm) but not losses of chromosome arms. Thus, chromosome changes occurring during tumor progression consist of chromosome gains. They involve the lq arm and chromosomes 10,2, 7, and 12, in decreasing order of involvement. This pathway of evolution is obviously different from that of breast and lung cancers, but recalls that observed in the trisomic type of colorectal cancers, in which chromosomes 7 and 12 were also frequently duplicated. As in this tumor type, endoreduplications were never observed. IV. Microsatellite instability in Epithelial Tumors

In colorectal cancer, we have seen that RER+ parallels the trisomic/normal tumor type and RER- phenotype the monosomic tumor type. To our knowledge, such a dual study between cytogenetic pattern and microsatellite instability has not been performed on other tumors. However, for breast and lung carcinomas, being all or almost all of the rnonosomic type, one should expect that the RER+ phenotype is rare in these tumors; in the hypothesis RER phenotype and chromosome alterations are always correlated. T h e studies on RER phenotype for these two tumor types remain quite limited. A. BREASTCANCER In breast cancer, two studies report quite heterogeneous results. A series of 104 tumors was studied by Wooster et al. (1994) at 12 loci. Abnormalities in microsatellite repeats were detected in 11 tumors, but only one anomaly was found in any single tumor. Thus, if microsatellite instability exists in breast cancer, it is characterized by a pattern different from that observed in colorectal cancer, where several loci were affected in tumors with the RER+ phenotype. Thus, depending on the definition given for the RER+ phenotype, 0-1076 of breast cancers may be regarded as RER+ . T h e series studied by Yee et al. (1994) was limited to 20 (lases. They found four cases (20%) with microsatellite instability and two cases having more than two loci affected. Thus, the two studies are quite contradictory, and more data are necessary to form a conclusion. However, even if the series of Yee et al. (1994) is not representative, it suggests that a small percentage of breast cancers may be RER+ . From the karyotypic data given earlier, we might have expected that no or almost no cases would be RER+, according to the definition given for colorectal

CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS

75

cancers. However, it must be recalled that, in the series we studied for cytogenetics, a proportion of breast cancers did not grow in vitro and thus failed to give information. Furthermore, only normal metaphases were observed in about 20% of the cases, which may be due largely to the analysis of noncancerous cells, but a proportion may also correspond to cancerous cells with normal chromosomes. At least one case of breast cancer with normal chromosomes was demonstrated as being tumorigenic by xenografting on nude mice (Gioanni et al., 1990). The high proportion of cells with normal or slightly rearranged karyotypes described by Bardi et al. (1993) is indeed highly suggestive of the cellular heterogeneity of breast tumors, but can hardly be related to cancer cells in the absence of a demonstration of their malignant or tumorigenic phenotype. Nevertheless, it remains possible that a (small?) proportion of breast cancers do not follow the monosomic type evolution and that these cases constitute a separate group of breast cancers. In their study, Yee et al. ( I 994) noticed that the tumors with a RER+ phenotype were of a low hiistological grade. In contrast, those with allelic losses (LOH) were not RER+, but had a higher histological grade. They concluded that microsatellite instability was an early and LOH a late occurring event of the tumor progression. This is likely to be true, but the two events are hardly sequential. An interpretation would be that RER+ tumors constitute a srnall subset of tumors with a slow evolution, keeping normal or fairly normal karyotypes and probably a good prognosis. Tumors with a RER- phenotype would strongly rearrange their karyotypes and have an adverse prognosis. Their evolution would tend to accumulate deletions, unmasking recessive mutations of tumor suppressor genes as in nnonosomic type colorectal cancers. Data on the karyotypes of breast tumors from predisposed patients are lacking. In our series, a number of patients had a number of phenotypic characteristics evoking a predisposition (young age, familial recurrence, multifocality), but no molecular confirmation of their predisposition. In contrast with colorectal cancers, where tumors from H NPCC patients have normal karyotypes, their tumors were among those with the most rearranged karyotypes. This observation fits with the absence of breast cancers among HNPCC families and the observation that genetic predisposition to breast cancer is determined by alterations of other genes, such as BRCAl or TP53 (Malkin et al., 1990; Hall et al., 1990), rather than mismatch repair genes. In breast cancer, the RER+ phenotype, if it exists, would be determined by somatic mutations.

76

BERNARD DUTRILLAUX

B. LUNGCANCER In lung cancer, a frequent microsatellite instability is described in both small-cell and non-small-cell (NSCLC) cancers with apparently high frequencies, 0.3-0.45 (Shridhar et al., 1994; Merlo et al., 1994). In most cases, this instability was limited to one or two loci, with several loci being altered in 6/38 (16%)NSCLC carcinomas only, an occurrence similar to that of colorectal cancers. Since all or almost all lung cancers seem to have deeply rearranged karyotypes, a correlation between chromosome and microsatellite instability data can hardly be proposed. C. ENDOMETRIAL CANCER Microsatellite instability was studied in a series of 30 endometrial cancers. It was observed at several loci ( 2 3 / 7 studied) in 7/30 (23%)cases (Tucker Burks ~t al., 1994). This rate is fairly high, possibly higher than that in other tumors in general. In the hypothesis that both normal and trisomic type karyotypes strictly correlate with RER+ phenotype, one should expect all endometrial cancers to be RER+, which is not the case. However, tumors with only normal or balanced karyotypes represent about 20% of the cases, which may fit with the frequency of RER+ tumors. It is noteworthy that endometrial carcinomas occur in women from HNPCC families and that this tumor is the most frequent extracolonic malignancy in these families (Lynch et al., 1993). Thus, the high occurrence of colorectal cancers with normal karyotype in HNPCC families suggests that quite similar oncogenic pathways exist in colorectal and endometrial cancers with normal karyotypes. Dual cytogenetic and molecular studies would provide interesting information. In conclusion, microsatellite instability may occur in various proportions of epithelial cancers, but its frequency of occurrence remains uncertain for several reasons. T h e number of tumors studied is still limited for each tumor type, the criteria used to define this instability vary, and different mechanisms may be a cause, varying according to the tumor type. A rough estimate would be that 0-25% of tumors are of the RER+ phenotype. A relationship between genetic predisposition and RER+ phenotype is documented in colorectal cancer, but not for other cancers where there is no indication that such a relationship may exist, except for endometrial carcinomas. A relationship between RER+ phenotype and a lack of chromosome deletions (tumors with normal or trisomic type karyotypes) is now documented in colorectal cancer, where the presence of a normal karyotype is suggestive of genetic predisposition.

CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS Germinal

mutation "

Somatic mutation

77

Tumor progression

.

'

Tumor suppressor gene

Few other recessive mutations unmasked by deletions

Y

Mismatch repair gene Many mutations, including dominant ones, expressed without deletion

FIG. 4. Hypothetical scheme representing two patterns of karyotype evolution in tumors from predisposed patients. Top: The recessive mutation of a tumor suppressor gene is Followed by multiple chromosome deletions during tumor progression. Bottom: The recessive mutation of a mismatch repair gene is followed by a passage to homozygosity. At this stage, the cell has no malignant character, but acquires a strong genomic instability. Arnong the multiple resulting mutations, some may dominantly activate oncogenic functions without the need for deletions during tumor progression.

Whether or not this can be extrapolated to other cancers remains to be determined. As a working hypothesis (Fig. 4),it may be proposed that other genetic predispositions are associated with the monosomic type pathway, in which many deletions occur. This would be the case, particularly for mutations of tumor suppressor genes such as that of p53 for breast cancer and APC for colorectal cancer. V. Concluding Remarks

Chromosome alterations in epithelial tumors are multiple. Their high complexity and their variability within a tumor type and from tumor type to tumor type have masked a number of common features that can

78

BERNARD DUTRILLAUX

now be highlighted. To reach such goal, it is, however, necessary to consider that very distinct processes of chromosome alteration exist. The most frequent, called monosomic, consists of accumulating unbalanced chromosome rearrangements, leading to deletions. Some are common to almost all tumor types, such as the deletions of lp, 8p, and 17p arms, whereas others are restricted to one o r a few tumor types, such as 5q in colorectal, 6q and 13 in breast and lung, and 9p in lung cancers, among others. Each tumor can be characterized by a pattern of deletions, but not by a single anomaly. These deletions are associated with whole chromosome losses and frequent endoreduplications. They may lead to the hemizygosity of up to 30% of the whole genome, as in hypodiploid breast cancers. This suggests that multiple recessive mutations can play a role, particularly when they involve tumor suppressor genes. Each tumor would thus be characterized by the concerted loss of a number of tumor suppressors. Some genetic predispositions may be prone to induce this pathway of evolution. This may be the case for the APC gene in colorectal cancer and the TP53 gene in breast cancer. Another pathway, called trisomic, consists of accumulating chromosome gains, whereas few or no unbalanced rearrangements leading to deletions occur. It is observed in 20-25% of colorectal and in a majority of endometrial carcinomas. In colorectal cancer, a high frequency of microsatellite instability was detected in this pathway, which can also be differentiated from the monosomic one by its metabolism and preferential proximal localization along the digestive tract. A third type of tumor is characterized by a high stability of the karyotypic, which remains normal. Its frequency remains unknown in most tumors, because the observation of a normal karyotype in a tumor may indicate that noncancerous cells were analyzed. Its reality, however, was demonstrated by tumorigenicity assays in colorectal cancers and in a single case of breast cancer. In colorectal cancer, this form with normal chromosomes is associated with HNPCC predisposition and microsatellite instability. Whether or not this opposition between microsatellite instability and chromosome stability is a characteristic shared by other tumor types remains an open question, but it constitutes a heuristic hypothesis with which to approach the mechanisms of genome alterations of cancer cells. REFERENCES

Aaltonen, L. A . Peltornaki, P., Leach, F. S., Sistonen, P., Pylkkanen, L., Mecklin, J. P., Jarvinen, H . , Powell, S. M., Jen, J., Hamilton, S. R., Petersen, G. M., Kinzler, K. W., Vogelstein, B., and d e la Chapelle, A. (1993). Sczence 260, 812-816.

CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS

79

Aaltonen, L. A., Peltomaki, P., Mecklin, J. P., Jarvinen, H., Jass, J. R., Green, J. S., Lynch, H. T., Watson, P., Tallqvist, G., Juhola, M., Sistonen, P., Hamilton, S. R., Kinder, K. W., Vogelstein, B., and de la Chapelle, A. (1994) Cancer Res. 54, 1645-1648. Almeida, A., Kokalj-Vokac, N., Lefranqois, D., Viegas-Pequignot, E., Jeanpierre, M., Dutrillaux, B., and Malfoy, B. (1993). Hum. Genet. 91, 538-546. Baker, S. J., Fearon, E. R., Nigro, J. M., Hamilton, S. R., Preisinger, A. C., Jessup, J. M., vanTuinen, P., Ledbetter, D. H., Barker, D. F., Nakamura, Y., White, R., and Vogelstein, B. (1989). Science 244, 217-220. Bardi, G., Johansson, B., Pandis, N., Mandahl, N., Bak-Jensen, E., Lindstrom, C., Tornqvist, A., Frederiksen, H., Andren-Sandberg, A., Mitelman, F., and Heim, S. (1993).Znt. J. Cancer 55, 422-428. Bardot, V., Luccioni, C., Lefranqois, D., Muleris, M., and Dutrillaux, B. (1991). Znt. J. Cancer 47,670-674. Bodmer, W. F., Bailey, C. J., Bodmer, J., Bussey, H. J. R., Ellis, A., Gorman, P., Lucibello, F. C., Murday, V. A,, Rider, S. H., Scambler, P., Sheer, D., Solomon, E., and Spurr, N. K. (1987). Nature 328, 614-616. Bravard, A., Luccioni, C., Muleris, M., Lefrancois, D., and Dutrillaux, B. (1991). Cancer Genet. Cytogenet. 56, 45-56. Bravard, A,, Hoffschir, F., Sabatier, L.,Ricoul, M., Pinton, A., Cassingena, R., Estrade, S., Luccioni, C., and Dutrillaux, B. (1992). Znt. J. Cancer 52, 797-801. Broccoli, D., and Cooke, H. (1993). Am. J. Hum. Genet. 52,657-660. Bronner, C. E., Baker, S. M., Morrison, P. T., Warren, G., Smith, L. G., Lescoe, M. K., Kane, M., Earabino, C., Lipford, J., Lindblom, A., Tannergard, P., Bollag, R. J., Godwin, A. R., Ward, D. C., Nordenskjold, M., Fishel, R., Kolodner, R., and Liskay, R. M. (1994). Nature 368, 258-261. Cavenee, W. K., Dryja, T. P., Phillips, R. A., Benedict, W. F., Godbout, R., Gallie, B. L., Murphree, A. L., Strong, L. C., and White, R. L. (1983).Nature (London)305,779-784. Chaum, E., Ellsworth, R. M., Abramson, D. H., Haik, B. G., Kitchin, F. D., and Chaganti, R. S. K. (1984). Cytogenet. Cell Genet. 38, 82-91. Couturier, J., Vielh, P., Salmon, R., and Dutrillaux, B. (1986).Znt. J. Cancer 38, 17-19. Couturier, J., Vielh, P., Salmon, R. J., Lombard, M., and Dutrillaux, B. (1988). Cancer Genet. Cytogenet. 33, 67-76. Couturier-Turpin, M. H., Couturier, D., Nepveux, P., Louvel, A., Chapuis, Y., and Guerre, J. (1982). Br. J. Cancer 46, 856-868. Delattre, O., Law, D. J., Remvikos, Y., Sastre, X., Feinberg, A. P., Olschwang, S., Melot, T., Salmon, R. J., Validire, P., and Thomas, G. (1989) Lancet, 353-355. Dutrillaux, B., and Couturier, J. (1986). Ann. Genet. 29, 76-81. Dutrillaux, B., and Muleris, M. (1986).Ann. Genet. 29, 11-15. Dutrillaux, B., Gerbault-Seureau, M., and Zafrani, B. (1990). Cancer Genet. Cytogenet. 49, 203-217. Dutrillaux, B., Gerbault-Seureau, M., Remvikos, Y., Zafrani, B., and Prieur, M. (1991). Breost Cancer Res. Treat. 19, 245-255. Fearon, E. R., Cho, K. R., Nigro, J. M., Kern, S. E., Simons,J. W., Ruppert, J. M., Hamilton, S. R., Preisinger, A. C., Thomas, G., Kinder, K. W., and Vogelstein, B. (1990). Science 247,49-56. Fliiry-HCrard, A., Viegas-Pequignot, E., De Cremoux, H., Chlecq, C., Bignon, J., and Dutrillaux, B. (1992). Cancer Genet. Cytogenet. 59, 1-8. Fujita, H., Wake, N., Kutsuzawa, T., Ichinoe, K., Hreshchyshyn, M. M., and Sandberg, A. A. (1985). Cancer Genet. Cytogenet. 18, 283-293. Giibas, Z., and Ruhin, S. C. (1987). Cancer Genet. Cytogenet. 25, 21-26.

80

BERNARD DUTRILLAUX

Gioanni, J., Le Francois, D., Zanghellini, E., Mazeau, C., Ettore, F., Lambert, J. C., Schneider, M., and Dutrillaux, B. (1990).Br.1. Cancer 62, 8-13. Groden, J., Thliveris, A., Samowitz, W., Carlson, M., Gelbert, L., Alhertsen, H., Joslyn, G., Stevens, J.. Spirio, L., Robertson, M., Sargeant, L., Krapcho, K., Wolff, E., Burt, R., Hughes, J. P., Warrington, J., McPherson, J., Wasmuth, J., Le Paslier, D., Abderrahim, H., Cohen, D., Leppert, M., and White, R. (1991). Cell 66, 589-600. Hall, J. M., Lee, M. K., Newman, B., Morrow, J. E., Anderson, L. A,, Huey, B., and King, M. C. (1990). Science 250, 1684-1689. Hartwell, L. (1992). Cell 71, 543-546. Hayashi, Y., Kanda, N., Inaba, T., Hanada, R., Nagahara, N., Muchi, H., and Yamamoto, K. (1989). Cancer 63, 126-132. Hoffschir, F., I’uillaume, M., Sabatier, L., Ricoul, M., Daya-Grosjean, L., Estrade, S., Cassingena, R., Cahzayrac, R., Sarasin, A.. and Dutrillaux, B. ( 1 993). Carcinogenesis 14, 1569- 1572. Huher, H., Knogler, W., Karlic, H., .4krad, M., Soregi, G., and Schweizer, D. (1990). Cancer Genet. Cy:ylogetwt.50, 189- 197. Jenkyn. D. J., and McCarttiey, A. J. (1986). Cancer Genet. Cytogenet. 20, 149-157. Kallioniemi, A., Kallioniemi. 0. P., Piper, J.,Tanner, M., Stokke, T.,Chen, L., Smith, H. S., Pinkel, D., Gray. J. W., and Waldman, F. M. (1994). Proc. Natl. Arud. Sci. USA 91,21562160. Kaneko, Y., Kanda, N., Maseki, N., Sakurai, M., Tsuchida, Y., Takeda, T., Okabe, I., arid Sakurai, M. (1987). Cancer RcJ. 47, 311-318. Kokalj-Vokac, N.. Almeida, A , , Viegas-Pequignot, E.. Jeanpierre, M., Malfoy, B., and Dutrillaux, B. (1993). Cjtogeizet. Cell Genet. 63, 11-15, Lasserre, C.. Sabatier, L., Beaumatin, J., Luccioni, C., LefranCois, D., Muleris, M., and Dutrillaux, B. (19942). Int. J . Cancer 56, 506-5 1 1. tasserre, C . , Sabatier, L., Salmon, R. J., and Dutrillaux, B. (1994h). In&.J . Oncol. 5 , 12671271. Lefranqlis, D., Olschwang, S., Delattre, 0..Muleris, M., Dutrillaux, A. M.. Thomas, G., and Dutrillaux, B. (1989). 1nt.J. Cancer 44,871-878. Levin, B., and Reirhmann, A. (1986).Caricer Genet. Cytogeriet. 19, 159-16‘2. Lukeis, R., Irving, L., Garson, M., and Hasthorpe, S. (1990). G ~ n e sChrom. Cancer 2, 116124. Lynch, H. T., Srriprk, T. C.. Watson. P.. Lanspa, S. J., Lynch, J. F., Lynch, P. M., Cavalieri, R. J., and Boland, (3. R. (1993). Grutroentei-olog?’ 104, 1555-1549. Magdelenat, H.. Gerhault-Seureau, hf., Laine-Bidron, M.,Prieur, C., and Dutrillaux, B. (1992). Bread CnncerRe.5. Treat. 22, 119-127. Magdelenat, H . , Gerbault-Seureau. M., and Dutrillaux, B. (1994). f u t . J . Cancer 57, 63-66. Malkin, D., Li, F. P., Strong, L. C . , Frdumeni. J. F., Nelson, C. E., Kim, D. H., Kassel, J., Gryka, M. A., Bischoff, F. Z., Tainsk!; $1. A., and Friend, S. H. (1990). Science 250, 1233-1238. hlerlo. A , , Xfahry. M., Gabrielson, E., Vollmer. K., Baylin, S. B., and Sidransky, D. (1994). Caiirer RP.F.54, 2098-2 10 1. Muleris, M., Salmon, R. J., Zafrani, B., Girodet, J., and Dutrillaux. B. (1985). Ann. G e n e t . 28, 206-2 13. Muleris, M.,Salmon, R. J.. and Dutrillaux, B . (1988). Cancer Genet. Cytogpnet. 32, 43-50, Yuleris, M.. Delattre, O., Olschwang, S., Dutrillaux. A . M., Remvikos, Y., Salmon, R. J., Thomas, G.. and Dutrillaux. B. (1990a). Cmicet- Gerzet. Cytog~net.44, 107-1 18. hluleris, X l . , Salmon, R. J.. and Dutrillaux, B. (199Ob). Cnncer Genet. Cytogenrt. 46, 143-156.

CHROMOSOME ALTERATION IN HUMAN EPITHELIAL CANCERS

81

Muleris, M., Almeida, A., Gerbault-Seureau, M., Malfoy, B., and Dutrillaux, B. (1994). Genes Chrom. Cancer 10, 160-170. Muleris, M., Dutrillaux, A. M., Olschwang, S., Salmon, R. S., and Dutrillaux, B. (1995). Genes Chrom. Cancer, in press. Nagase, H., and Nakamura, Y. (1993). Hum. Mutat. 2,425-434. Nicolaides, N. C., Papadopoulos, N., Liu, B., Wei, Y. F., Carter, K. C., Ruben, S. M., Rosen, C. A., Haseltine, W. A., Fleischmann, R. D., Fraser, C. M., Adams, M. D., Venter, J. C., Dunlop, M. G., Hamilton, S. R., Petersen, G. M., de la Chapelle, A., Vogelstein, B., and Kinzler, K. W. (1994). Nature 371, 75-80. Ochi, H., Takeuchi, J., Holyoke, D., and Sandberg, A. (1983). Cancer Genet. Cytogenet. 10, 12 1-122.

Paridis, N., Heim, S., Bardi, G., Idvall, I., Mandahl, N., and Mitelman, F. (1993). Genes Chrom. Cancer 6, 51-57. Partdis, N., Bardi, G., and Heim, S. (1994). Cancer Genet. Cytogenet. 76, 77-84. Papadopoulos, N., Nicolaides, N. C., Wei, Y. F., Ruben, S. M., Carter, K. C., Rosen, C. A,, Haseltine, W. A,, Fleischmann, R. D., Fraser, C. M., Adams, M. D., Venter, J. C., Hamilton, S. R., Petersen, G. M., Watson, P., Lynch, H. T., Peltomaki, P., Mecklin, J. P., d e la Chapelle, A,, Kinzler, K. W., and Vogelstein, B. (1994). Science 263, 1625-1628. Reichmann, A., Martin, P., and Levin, B. (1981). Int. J. Cancer 28, 431-440. Rernvikos, Y., Gerbault-Seureau, M., Vielh, P., Zafrani, B., Magdelenat, H., and Dutrillaux, B. (1988a). Cytometry 9, 612-618. Rernvikos, Y., Muleris, M., Vielh, P., Salmon, R. J., and Dutrillaux, B. (1988b). Znt.J. Cancer 42,539-543.

Rernvikos, Y., Gerbault-Seureau, M., Magdeltnat, H., Prieur, M., and Dutrillaux, B. (1992). Breast Cancer Res. Treat. 23, 43-49. Rernvikos, Y., Magdelenat, H., and Dutrillaux, B. (1995a). Breast Cancer Res. Treat., in press. Rernvikos, Y., Vogt, N., Muleris, M., Salmon, R. J., Malfoy, B., and Dutrillaux, B. (1995b). Genes Chrom. Cancer, 12, 272-276. Saint-Ruf, C., Gerbault-Seureau, M., Viegas-Ptquignot, E., Zafrani, B., Malfoy, B., and Dutrillaux, B. (1991). Cancer Genet. Cytogenet. 52, 27-35. Sartdberg, A. A. (1990). “The Chromosome in Human Cancer and Leukemia,” 2nd ed. Elsevier, New York, Amsterdam, Oxford. Schmitt, H., Blin, N., Zankl, H., and Scherthan, H. (1994). Genes Chrom. Cancer 11,-171177.

Schott, D. R., Ngo Chang, J., Deng, G., Kurisu, W., Kuo, W. L., Gray, J., and Smith, H. S. (1994). Cancer Res. 54, 1393-1396. Shah, N. K., Currie, J. L., Rosenshein, N., Campbell, J., Long, P., Abbas, F., and Griffin, C. A. (1994). Cancer Genet. Cytogenet. 73, 142-146. Shridhar, V., Siegfried, J., Hunt, J., del Mar Alonso, M., and Smith D. I. (1994). CancerRes. 54, 2084-2087.

Testa, J. R., Siegfried, J. M., Liu, Z., Hunt, J. D., Feder, M. M., Litwin, S., Zhou, J. Y., Taguchi, T., and Keller, S. M. (1994) Genes Chrom. Cancer 11, 178-194. Th,arapel, S. A,, Qumsiyeh, M. B., and Photopulos, G. (1991). Cancer Genet. Cytogenet. 55, 89-96.

Thibodeau, S. N., Bren, G., and Schaid, D. (1993). Science 260, 816-819. Tucker Burks, R., Kessis, T. D., Cho, K. R., and Hedrick, L. (1994). Oncogene 9, 1163-1 166. Viegas-PCquignot, E., Flury-Hkrard, A., De CrCmoux, H., Chlecq, C., Bignon, J., and Dutrillaux, B. (1990). Cancer Genet. Cytogenet. 49, 37-49. Weber, G., Kizaki, H., Tzeng, D., Shiotani, T., and Olah, E. (1978). Life Sci. 23, 729-736.

a2

BERNARD DUTRILLAUX

Wooster. R . , Cleton-Jansen, A . M., Collins, N.,Mangion, J., Cornelis, R. S., Cooper, S. C., Gusterson, B. A , , Ponder, B. A . J., von Deimling A . , Wiestler, 0. D., Cornelisse, C. J., Devilee, P., and Stratton, M. R. (1994). Nature Genet. 6, 152-156. Yee, C. J., Roodi, N . , Verrier, C. S., and Parl, F. F. (1994). Cancer Res. 54, 1641-1644. Yin. Y., Tdinsky, M. A , , Bischoff, F. Z., Strong, L. C., and Wahl, G. M . (1992). Cell 70,937948. Yoshida, M. .4., Ohyashiki, K., Piver, S. M., and Sandberg, A. A. (1986). Cancer Genet. Cytogmel. 20, 159- 162.

GENETICS OF MURINE LUNG TUMORS Tommaso A. Dragani, Giacomo Manenti, and Marco A. Pierotti Division of Experimental Oncology A, lstituto Nazionale Tumori, 20133 Milan, Italy

I. Historical Aspects 11. Comparative Aspects A. Histology B. Gene Mutations C. Gene Expression D. Loss of Heterozygosity (LOH) 111. Genetic Linkage Studies IV. Transgenic Models of Lung Tumorigenesis V. Candidate Lung Tumor Susceptibility Genes VI. Genetics of Lung Tumors, Human ’VII. Conclusions and Perspectives References

1. Historical Aspects

Inheritance of susceptibility to lung tumors in mice was first shown by Lynch as early as 1926 (Lynch, 1926). Since its establishment in 1921, the mouse inbred strain A has progressively become the model for most subsequent studies on lung tumor susceptibility. In the 1930s, Strong (1936) and Bittner (Bittner, 1938, 1939) reported the high spontaneous incidence of pulmonary tumors in this strain. Andervont (1937) showed that virtually all young mice of this strain developed multiple lung tumors within 2 months following the subcutaneous application of dibenz[a,h]anthracene. Subsequently, Heston carried out detailed genetic studies by crossing the A strain with different mouse strains. As a result, he reached the following conclusions: (1) lung tumor susceptibility is a partially dominant trait, since F 1 animals between susceptible and resistant strains are intermediately susceptible; (2) multiple genes are involved in determining lung tumor susceptibility. He was the first to suggest that this susceptibility behaves as a “quantitative character controlled by multiple factors” (Heston, 1940, 1942). By using different crosses, however, other authors suggested that a single gene determines the difference in susceptibility to lung tumors (Bittner, 1938; Andervont, 1940). Some studies have indicated that susceptibility to lung tumorigenesis 83 ADVANCES IN CANCER RESEARCH. VOL. 67

Copyright 0 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

a4

TOMMASO A . DRAGANI E T At..

is determined by a single gene, while other studies have suggested that multiple genes are involved. This discrepancy may again be due to the different strain combinations used (Bloom and Falconer, 1964; Malkinson and Beer, 1983; Malkinson et al., 1985). T h e mouse inbred strains that have been identified so far as highly susceptible to lung carcinogenesis comprise the A, SWR, 020, and NGP mice (Della Porta et al., 1967; Malkinson, 1989, 1991; Thaete et al., 1991). Most studies on the genetics of lung tumorigenesis in mice have considered tumor incidence and the number of tumors per animal as the phenotype, without taking into consideration the size of neoplastic lesions. However, we have proposed a quantitative analysis of genetic susceptibility that takes into account both the number of lung tumors and their volume (Dragani et al., 1991). There is no relationship between the susceptibility of any given mouse strain to lung tumors and its susceptibility to tumors of other organs. Susceptibility to spontaneous lung tumor development is paralieled by susceptibility to induction of the same tumor type by chemical carcinogens (Della Porta et al., 1967). Lung tumor induction in strain A has also been suggested as a mediuni-term bioassay system for carcinogenicity. T h e bioassay consists of an untreated control group of mice and groups that were administered with the test chemical at three dosage levels. Animals are observed for 16-24 weeks after treatment and then the lung tumors are counted (Shimkin and Stoner, 1975). Interestingly, all of the carcinogens found positive in the lung tumor bioassay are genotoxic (Pereira and Stoner, 1985; Maronpot et al., 1986). Mouse lung tumor assay has also been proposed as a screening system for developing new cancer chemotherapeutic agents. Intervention therapy is initiated after lung tumor induction by chemical carcinogen administration, when all mice have lung tumors. Cancer chemotherapeutic agents are then administered for a short period of time, and the number and size of lung tumors are scored in control and treated groups (Be1993). linsky et d., It. Comparative Aspects The study and identification of genetic factors affecting inherited predisposition to lung tumorigenesis in mice are of great interest as a model system for understanding pathogenetic mechanisms. At present, it is not clear whether these studies have applications for humans. Therefore, it is important to establish whether the mouse lung tumors represent the experimental counterpart of a human lung tumor histotype. Once the correspondence between mouse lung tumors and a par-

GENETICS OF MURINE LUNG TUMORS

85

ticular human lung tumor histotype has been established, it will be possible to transfer the results obtained in the experimental models to the human situation. A large portion of the mouse genome shows regions of homology and conserved syntenies with the human genome, and comparative genetic maps between mice and humans have been used to predict the location of human and murine disease genes on the basis of their mapping in other species (Lyon et al., 1990; Leff et al., 1992; Su et al., 1992; Copeland et al., 1993; Tassabehji et al., 1993; Levinson et d.,1994). The mapping of the genes responsible for genetic susceptibility and resistance to lung carcinogenesis in mouse chromosomes, and the subsequent analysis of homology between mice and humans on the chromosomal regions containing putative tumor susceptibility and resistance gene(s), would suggest that we should test genetic markers localized in the corresponding human chromosomal region for possible linkage with the risk of lung tumor in humans. The identification of the chromosomal localization of loci predisposing lung tumor development could be the first step toward the cloning and identification of these genes. Once the nnurine genes affecting lung tumor susceptibility have been identified and cloned, it will be possible to test their human homologues for the presence of mutations in human lung tumors and in the germ line of patients affected with a lung tumor. An excellent review on comparative aspects of mouse and human lung tumors indicates that mouse lung tumors represent the experimental model for human adenocarcinomas (Malkinson, 1992). Here, we will focus on some of the comparative aspects of the relationship with the genetics of lung tumor susceptibility and on some findings that were published after this previous review. A. HISTOLOGY

Histopathological characteristics of mouse lung tumors have already been described in detail (Stewart et al., 1979). It is interesting to note that both spontaneous and induced lung tumors are histologically similar. Elriefly, two histological types of lung tumors have frequently been reported. T h e solid o r alveolar form consists of oval or cuboidal cells that invest the alveoli or that produce cords or solid nests, and the papillary type consists of tumor cells arranged in papillary formations. However, mixed morphological patterns have also been observed, and the progression of the neoplastic lesions, from a solid to papillary growth pattern, has been reported (Grady and Stewart, 1940; Kimura, 1971; Stewart et d.,1979; Belinsky et al., 1992a). Strain differences have been reported

86

TOMMASO A. DRAGANI E T AL.

for the histological type of lung tumors induced by chemical carcinogens. Indeed, the relative proportion of solid and papillary adenomas varies between strains, with possible genetic dominance of the papillary phenotype (Beer and Malkinson, 1985; Thaete et al., 1987, 1991). The cellular origins of mouse lung tumors have been debated in the past. Major evidence indicates that they originate from alveolar type I1 celfs because they express common antigens and genes and have common ultrastructural characteristics with the tumors (Ward et al., 1985; Rehm et al., 1988; Belinsky et al., 1992a; Re et al., 1992). However, an origin from Clara cells of some tumors has also been suggested (Thaete et al., 1991). In humans, small-cell lung cancer (SCLC) and non-small-cell lung cancer (NSCLC) are the two main groups of lung tumors. NSCLCs account for about 75% of cancers and include three major histological subtypes: squamous cell carcinoma, adenocarcinoma, and large-cell carcinoma. All of these lung tumor types represent distinct morphological, biological, and clinical entities. The human lung tumor histotype that is closest to the mouse lung histotype is the lung adenocarcinoma. Microscopically, the latter tumors exhibit a wide range of differentiation, with one extreme resembling bronchioloalveolar carcinoma and the other resembling undifferentiated large-cell carcinoma. Lung adenocarcinomas are characterized by glandular differentiation, with the formation of tubules or papillae (Rosai, 1989).

B. GENEMLTATIONS I . KRAS2 Mutations

Changes in the structure of KRAS2 most commonly affect codons 12, 13, and 61, generating a permanently activated p2lras protein that contributes to the development of cancer [for a review, see Barbacid (1987)l. Several studies have identified Kras2 mutations in spontaneous or chemically induced lung tumors in mice. Table I summarizes the data available, indicating the percentage and the mutational spectrum affecting the Kras2 gene along with the strains and treatments used (see Table I). In spontaneous lung tumors of mice, the frequency of Kras2 mutations is usually very high. Indeed, lung tumors that arose in untreated mice of strains A/J and CD-1 and from hybrids (C3H/He x A/J)Fl and (BALBlc X DBA/2)F1 were affected in 77-95% of all cases (You et al., 1989, 1992a; Manam ef al., 1992; Herzog et al., 1993; Li et al., 1994a). However, the rate of activation in spontaneously occurring lung tumors from resistant strains was 43, 10, and 17% in C3H/HeJ, B6C3F1, and

GENETICS OF MURINE LUNG TUMORS

87

C!57BL/6J mice, respectively (Candrian et al., 1991; Devereux et al., 1091, 1993). Several carcinogens belonging to different chemical classes were used to induce lung tumors in both susceptible and resistant mice. In most reported cases, the rate of Kras2 mutation ranged from 80 to 100% (see Table I). In chemically induced lung tumors, the frequency of Krm2 activation was similar in resistant mice (C3H/HeJ, C57BL/6J, B6C3Fl) and in the susceptible strain A/J (Stowers et al., 1987; Goodrow et al., 1990; Devereux et al., 1991, 1993), suggesting that susceptibility to chemically induced lung carcinogenesis does not correlate with the mutability of the Krm2 gene. This conclusion is also supported by Fijneman et al. (1994a), who found different frequencies of Kras2 mutations in inbred and congenic inbred strains displaying similar susceptibility to lung carcinogenesis. If the mutation type is now considered, both codons 12 and 61 of Krm2 in spontaneous lung tumors contain activating mutations with no selectivity for a codon o r a specific base substitution (Table I) (You et al., 1!389, 1992a; Devereux at al., 1991; Manam et al., 1992; Li et al., 1994a), except for the results of Herzog et al. (1994) on (BALB/c x DBA)Fl mdce. In this study, lung tumors obtained from the vehicle-treated group resulted in mutation only at codon 61. However, chemically induced lung cancers show a mutation spectrum of Kras2 that appears to be specific for each chemical used and shows selective involvement of a codon and a preferential base substitution (see Table I). This is often consistent with the adduct profile of the specific chemical, although the final base substitution does not always parallel the concentration of a p(articu1ar adduct, but is probably influenced by other factors such as differences in repairing a specific lesion, the transforming potential of an amino acid change, or still unknown host- and tissue-specific factors (Der et al., 1986; Belinsky et al., 1989; Devereux et al., 1991, 1993; Ohmori et al., 1992; Mass et al., 1993; Wang et al., 1993; Li et al., 1994a; You et al., 1994). It is worth noting that the amount and the persistence of 06-methylguanine induced by 4-(N-methyl-N-nitrosamino)-1-(3pyridy1)-l-butanone (NNK) in the susceptible A/J and in resistant C57BL/6J mice d o not account for the difference in lung tumor susceptibility (Devereux et al., 1993). Krm2 activation in mice is an early event in the multistep process of lung carcinogenesis since it is invariably found in both adenomas and adenocarcinomas without any significant difference in the rate of mutation observed (Belinsky et al., 1989; You et al., 1989; Devereux et al., 1993; Herzog et al., 1993). The acquisition of Kras2 mutation during the very earliest phases of tumor formation was confirmed by Belinsky et al.

Nonc

!I/]

lO"/I I 9/,

[(jl%)

I

3 4

2 2

3

2 2

1

1

116 (17%)

AAF

17/20 (85%) 5 / 6 (83%) J/7 ( 4 9 4 ) 2/3 ( 6 7 % ) 1 / 1 0 (107) 2012 I (!15%) 20126 (77%) 212 (100%) 4/4 (lOO2) 1 o/ 12 (89%>)

I

Ulill' BTC Butadiene DMBA ENU

IQ MNU

I

1 9

I

2 1 1

I

1

9 4

15

I 1

I

4

I

4 1

2 4 5

14"/16 (88%)

22/27 (82%) 14/15 (935%) 15/17 (88%) 23/23 (100%) 24/24 ( 100%) 6/9 (67%) 17/17 (1005%) 1 1 / 1 1 (100%) 13/63 (21%) 0/20 (0%) 1/51 (2%) 49/54 (91%) 15/15 (100%)

I

I I

2 I

2 I5

8 I2 3 9

I 4 4 I 2

1

3 1

6 4 4 7

I

3

1

24 2

3

1

3

6 2

2 7

5

1

1

1

13 1

7 15

39

2

You rl (11. ( 1989) Mass el nl. ( 1993) Devetellx rt al. ( I I N Y ) Manam ~t (11. (1992) 1.i et u1. ( 1994a) Devereux v t 01. ( I99 I ) Ohniori ~t al. (1992) Cantlrian ~t al. (I9!l I ) You PI nl. (IY92a) Herzog et al. ( 1993) Wdrlg f!t (11. (1993) Wang et (11. (1993) Li et nl. (IY94a) You et al. (1989) Mass et al. (1993) Li et al. (lW4a) Chen el al. (1994b) Mass et al. (1993) You et ul. (1993) Goodrow et al. ( 1990) Manam et ul. (1992) You et al. (1992b) Fijneman et al. (1 994a) Fijneman et al. ( 1 994a) Fijneman et al. (1994a) Herzog et al. (1993) You et al. (1989)

5-MC NDEA NDMA

6-NC NNK

cc

TNM Urethane

VC

44/49 (90%) 23/28 (82%) 6/16 (38%) 9/10 (90%) 11/11 (100%) 15/16 (94%) 13/14 (93%) 10111 (90%) 19/19 (100%) 46/59 (78%) C3H/He 11/11 (100%) C57BL/6J 2/22 (9%) ACSFI 19/20 (95%) 717 (100%) C3AFl 9 / 9 (100%) B6C3F1 10111 (91%) A/J 9/9d (100%) BALBlc 415 (80%) ACJFI 9/11 (82%)' B6AFl 24/25 (96%) AC3Fl 11/12 (92%) 13/14 (93%) C3AFl C57BL/61 - 11f/20 (55%)

A/J CD- 1 A/J C3H/He C3AFl CD- 1

12

22 5

10 2 1

16 5 3

6 11 14

5

1

1

4

7 19 46 11

1

1

2

1 2

1

19

7 9 1

2

2 4 3 2 11 5 6 3

7 3

1

1

4 11 6 7 4

1 1

1

2

1

You et al. (1994) You et al. (lYY2b) Manam et al. (1992) Belinsky et al. (1989) Devereux et al. (1991) Chen et al. (1994b) Li et al. (1 994a) Belinsky et al. (1989) Matzinger et al. (1994) Chen et al. (1993) Devereux et al. (1991) Devereux et al. (1993) You et al. (1992a) You et al. (1992a) Stowers et al. (1987) You et al. (1989) Re et al. (1992) Re et al. (1992) Re et al. (1992) Ohmori et al. (1992) You et al. (1992a) You et al. (1992a) Devereux et al. (1993)

.Abbreviations: AAF, 2-acetylaminofluorene; 6-AC, 6-aminochrysene; B[a]P, benzo[a]pyrene; Bb]P, benzo[i]pyrene; BTC, benzotrichloride; DMBA, 7,12dimethylbenz[a]anthracene;ENU, N-ethyl-N-nitrosourea; IQ, 2-amino-3-methylimidazo[4,5-f]quinoline; MNU, N-methyl-N-nitrosourea; 5-MC, 5-rnethylchrysene; I-(3-pyridyl)-1-butanone; TNM, tetraNDEA, N-nitrosodiethylamine; NDMA, N-nitrosodimethylamine; 6-NC, 6-nitrochrysene; NNK, 4-(N-methyl-N-nitrosamino)nitromethane; VC, vinyl carbamate. Hybrids: AC3F1, (A X C3H)Fl; B6AF1, (C57BL/6 X A)Fl; B6C3F1, (C57BL6/J X C3H)FI; CDFI, (BALB/c X DBAI2)Fl. bFor one sample the activating mutation was unknown. .The number of samples analyzed is not reported. d o n e tumor contained an undetermined mutation at the third base of codon 61. rOne sample was a pool of two small tumors and contained two different mutations. Three tumors contained an undetermined mutation at the third base of codon 6 1. /One adenocarcinoma had two different mutations.

90

TOMMASO A . DRAGANI E T AL.

(1992b), who detected a mutated Kras2 in 85% of hyperplastic lesions from NNK-treated A/J mice. T h e quality of the Krm2 mutations induced by urethane, 7,12-dimethylbenz[a]anthracene(DMBA), and nitrosodiethylamine (NDEA) seems to affect tumor progression. In fact, the transversion A+T at the second base of codon 6 1 was found prevalently in adenomas, while the transition A - 4 at the same base was present in adenocarcinomas (Nuzum et al., 1990; Manam et al., 1992; Ohmori et a!., 1992). In humans, adenocarcinomas of the lung have been analyzed extensively for activating mutations in KRAS2. At least 187 (23%) mutated adenocarcinomas out of 829 investigated cases (including our unpublished 42 mutated samples from 169 adenocarcinomas) have been described (Kobayashi et al., 1990; Mitsudomi et al., 1991; Reynolds et ul., 1991; Li Lung et al., 1992; Rodenhuis and Slebos, 1992; Sugio et al., 1992; Husgafvel-Pursiainen et al., 1993; Rosell et al., 1993; Westra et al., 1993; Kern et al., 1994; Li et al., 1994b). T h e involvement of KRAS2 is not reported in SCLC (Minna, 1993) and is uncommon in NSCLC subtypes other than adenocarcinomas (Mitsudomi et al., 1991; Reynolds et ul., 1991; Li Lung et al., 1992; Rodenhuis and Slebos, 1992; Sugio et al., I992; Husgafvel-Pursiainen et al., 1993). However, Rosell et al. (1993) found 8 of 38 (2 1%) cases of squamous cell carcinomas were activated. Most of the activating mutations in human adenocarcinomas occur at the 12th codon of KRAS2 (Rodenhuis and Slebos, 1992; Sugio et al., 1992),especially from a transversion G+T in the first two bases (Li Lung et al., 1992). Adenocarcinomas from smokers have a higher percentage of KRAS2 mutations than adenocarcinomas from nonsmokers (Slebos et al., 1991; Li Lung et al., 1992; Rodenhuis and Slebos, 1992; Sugio et al., 1992; Husgafvel-Pursiainen et al., 1993; Westra et al., 1993). Interestingly, the mutation W T found in smokers is also a typical mutation detected in benzo[a]pyrene (B[a]P) and 5-methylchrysene (5-MC) induced lung cancer in mice (You et al., 1989; Mass et al., 1993; Chen et al., 1994b; Li at at., 1994a), two compounds that are present in smoke (Loeb et al., 1984). KRAS2 mutation is indicated as an early event in the pathogenesis of human adenocarcinoma of the lung (Rodenhuis et al.,1987; Westra et al., 1993; Gazdar 1994: Li et al., 1994b) and represents a genetic marker for the identification of a subgroup of patients with poor prognosis (Slebos rt al., 1990; Mitsudomi et al., 1991; Rodenhuis and Slebos, 1992; Sugio et al., 1992). Putting all of these observations together would suggest that KRAS2 activation plays a fundamental role in the pathogenesis of lung adenocarcinoma in mice and in humans. In fact, this lesion is a frequent and

GENETICS OF MURINE LUNG TUMORS

91

early genetic alteration present in spontaneous and chemically induced lung tumors in mice, as well as in human lung adenocarcinomas. Furthermore, it may influence tumor progression in mice and is a prognostic marker in human lung adenocarcinamas. However, Kras2 activation does not distinguish murine strains that are susceptible from those that are resistant.

2. p53 Mutations Alterations affecting the p53 gene are one of the most common lesions in human tumors, including all types of lung cancer (Yokota and Sugimura, 1993). Five different mechanisms for p53 inactivation are outlined (Vogelstein and Kinzler, 1992). Among these, non-sense and splicing site mutations, producing a truncated p53 protein, or missense mutations, causing a change in the amino acid encoded, are fairly common in SCLCs and NSCLCs (Minna, 1993). In human lung adenocarcinomas, p53 mutations are reported with a frequency that ranges from 12 to 41%, with an average rate of 34% (Chiba et al., 1990; Kishimoto et al., 1992; Miller et al., 1992; Horio et at., 1993; Li et al., 1994b; Ryberg et al., 1994). Furthermore, the incidence of altered p53 has been suggested to be an early event and is a potential prognostic factor (Quinlan et al., 1992; Sozzi et al., 1992; Horio et al., 1993; Li et al., 1994b). These findings suggest that alterations on p53 play a significant role in the development of human adenocarcinomas. The presence of mutations on murine p53 gene (TTp53)was studied in exons 5-8, in lung tumors induced by different carcinogens [DMBA, NDEA, NNK, 6-nitrochrysene (6-NC), and vinyl carbamate (VC)], and in several murine strains (A/J, CD-1, and C57BL/6J) (Goodrow et al., 1992; Chen et al., 1993; Devereux etal., 1993; Li et al., 1994a). In none of these studies were mutations affecting this gene found. However, Hegi et al. examined 54 methylene chloride-induced lung tumors in B6C3F1 mice for p53 alterations. They found four cases of missense mutations among seven adenocarcinomas, which also lost heterozygosity for markers around the Trp53 gene (He@ et al., 1993). C. GENEEXPRESSION Despite the conspicuous amount of information on gene expression in human and murine lung canters, most of the work done on these two species is not easily comparable. Indeed, some aspects were investigated in only one model, making comparison impractical [see reviews in Buchhagen (1991), Malkinson (1992), and Minna (1993)l. Nevertheless, a direct analysis of similarities and differences for the expression of

92

TOMMASO A. DRAGANI E T A L .

surfactant proteins, Clara cell 10-kDa antigen (CClO), RB, p53, and c-mycis still feasible. Pulmonary surfactant is a complex mixture containing lipids, phospholipids, and four different surfactant proteins designated SP-A, SP-B, SP-C, and SP-D. All of these proteins are expressed selectively in epithelial cells of the lung and are synthesized by alveolar type I1 cells and Clara cells, with the exclusion of SP-C (Kuroki and Voelker, 1994). T h e expression of this protein appears to be confined to alveolar type I1 cells in mice (Wikenheiser et al., 1992; Kuroki and Voelker, 1994), but it seems to be present in both cell types in humans (Wikenheiser et al., 1992). SP-A expression has frequently been reported in human pulmonary adenocarcinomas, although at various incidences (Broers et al., 1992; Linnoila et al., 1992; Shijubo et al., 1992; Smith et al., 1994). By contrast, only a minority of other NSCLC subtypes express SP-A (Broers ut al., 1992; Linnoila et al., 1992). SP-B and SP-C were found, either constitutively o r after dexamethasone induction, in certain human lung cancer cell lines (Gazdar et al., 1990). In naturally occurring or chemically induced murine lung tumors, SP-A was found in most of the cases reported (Ward et al., 1985; Rehm et al., 1988; Re et al., 1992). High levels of SP-C mRNA were present in some lung tumors from transgenic mice with the SV40 large T antigen under the control of the human SPC promoting region (Wikenheiser et al., 1992). Some of these previous studies also analyzed the expression of CC10, which was used as a distinct marker for Clara cells (Singh et al., 1988) in pulmonary neoplasms of mice and in human adenocarcinomas. All but the tumors from transgenic mice (Wikenheiser et al., 1992) were negative for this marker (Ward et ul., 1985; Rehm et al., 1988; Broers et al., 1992). In order to further characterize the cells comprising the human and murine lung tumors, we studied the mRNA levels of SP-A, SP-B, and SP-C in a group of urethane-induced lung cancers from AC3F2 mice and in paired nonneoplastic and neoplastic lung tissues from 20 human lung adenocarcinomas. We invariably found expression of SP-A and SP-B and the absence of detectable CC10. Interestingly, SP-C, which is considered a specific marker for type I1 epithelial cells in mice, was missing in murine tumors, but was detectable in 3 of 20 human adenocarcinomas (De Gregorio et nl., manuscript in preparation). In our earlier study on mRIVA expression of several genes in urethane-induced lung tumors, we reported a strong reduction in the level of RB gene transcript (Re et al., 1992). This finding is also a common feature in human lung adenocarcinomas that show the loss of RB biochemical functions, not only by gene mutation or deletion but also by altered control of RB gene expression (Xu et a/.,1991; Minna, 1993;

GENETICS OF MURINE LUNG TUMORS

93

Reissmann et al., 1993; Xu et al., 1994). We have also reported increased c-myc mRNA levels in our lung tumor samples (Re et al., 1992), which is a common alteration found in the human counterpart (Gosney et al., 1990; Volm et al., 1993; Wodrich and Volm, 1993).

D. Loss

OF

HETEROZYGOSITY (LOH)

A few studies have reported LOHs in mouse lung tumors. Each of these studies took into account only one or a few chromosomes. On the other hand, a considerable amount of information has been published on cytogenetic changes and LOHs in human lung tumors of different histotypes. By comparative genome analysis of homologous chromosomal regions, it might be possible theoretically to distinguish LOHs occurring at the corresponding chromosomal regions in both species. These LOHs therefore would be important for tumor pathogenesis, compared to LOHs occurring in only one species, and so are probably of limited relevance to tumor development. Moreover, LOHs occurring at regions of homology in lung tumors of the two species (mouse and human) would strongly support a common pathogenetic mechanism. LOHs on Chromosome 4 were studied in a total of 147 mouse lung tumors induced by different chemical carcinogens or spontaneously developed in the (BALB/c X DBA/ZJ)Fl, (C3H/HeJ X A/J)Fl, and (A/J x C57BL/6J)F1 hybrid mice. LOHs were found with the highest frequency around the marker D4Mit77, localized near the a-interferon locus (Zfa).LOHs were detected in 29/61 (48%) of the lung adenocarcinomas but in only 1/38 (3%)of the lung adenomas examined, suggesting that these lesions contribute to the progression of mouse lung tumors. T h e authors did not examine other markers localized on other chromosomes (Herzog et al., 1994). LOHs in the same region on Chromosome 4 were reported by Wiseman et al. (1994) in 618 (75%) mouse lung tumors induced by butadiene in B6C3F1 mice. On Chromosome 11, Wiseman et al. (1994) found LOHs on loci surrounding the p53 tumor suppressor gene (Trp53) in 2 of 8 butadieneinduced B6C3F1 lung tumors. The same markers around Trp53 were studied for possible LOHs in 54 methylene chloride-induced and in 7 spontaneous lung tumors from B6C3F1 mice. LOHs were detected in seven methylene chloride-induced lung carcinomas (i.e., in 13% of the examined tumors) (Hegi et al., 1993). LOHs in the region of the retinoblastoma tumor suppressor gene ( R b l )on Chromosome 14 were infrequent in mouse lung tumors, as two studies reported LOHs at the Rbl locus in 3/61 and 1/8 tumors, respectively (Hegi et al., 1993; Wiseman et al., 1994).

94

TOMMASO A . DRAGANI E T A L .

Taken together, the three studies available on LOHs in mouse lung tumors showed frequent involvement of loci around Ifa on Chromosome 4 and rare involvement of Trp53 (Chromosome 11) and Rbl (Chromosome 14) loci. The critical region on Chromosome 4 implicated in LOH is homologous to human chromosome 9p21-22, where the human a-interferon (Ih’FA) locus is localized (Copeland et al., 1993; MGD, 1994). The same region undergoes frequent deletions in human primary tumors of different types, including lung tumors (Olopade f t al., 1993; Merlo ~t al., 1994). In human lung tumors, genomic alterations, including chromosome aberrations and LOHs, are massive (Lukeis et al., 1990; Buchhagen, 1991; Wang-Peng et al., 1991; Minna, 1993; Testa et al., 1994). A number of recurrent LOHs, occurring at the location of known or suspected tumor suppressor genes, have been demonstrated. At present, it is not known whether these multiple alterations reflect the multistep nature of lung tumor pathogenesis. Alternatively, the multiple cytogenetic and molecular changes may be a consequence of the altered function of a set of genes that controls genomic stability and plays a fundamental role in lung cancer pathogenesis. T h e pattern of LOHs at multiple chromosomal locations is a particular and recurrent characteristic of human lung tumors, irrespective of their histotype. LOHs at specific chromosomal regions are seen more frequently in a certain lung cancer histotype than in others, with no absolute histotype specificity. Here, we have restricted our review to studies carried out on human Chromosomes 9, 13q, and 17p, which represent the regions of homology to the mouse chromosomal regions where LOHs have already been found. On Chromosome 9p, LOHs occurred in 36-63% of lung adenocarcinomas (Merlo ct ul., 1994; Sato et al., 1994). Additional studies of lung cancer cell lines defined the minimal region of loss at 9p21-22, close to the IFN gene cluster (Olopade et al., 1993; Mead et al., 1994; Merlo et al., 1994). Recently, a putative tumor suppressor gene (named MTS1 or CDKNB), an inhibitor of a cyclin-dependent kinase, has been identified in the region 9p21 and found mutated in the germ line of melanoma patients (Hussussian rt al., 1994; Kamb et al., 1994). Mutations at the MTSl/CDKN2 gene have been found at low frequency in human primary lung tumors (Cairns et al., 1994; Okamoto et al., 1995). LOHs on Chromosome 13q, affecting the RB gene, have been found at a relatively high frequency in human lung adenocarcinomas (from 18 to 5 1?& of cases), as well as in other histotypes, with an incidence higher than 80% in SCLCs. In this tumor histotype, the remaining RB allele was

GENETICS OF MURINE LUNG TUMORS

95

often mutated (Yokota et al., 1987; Harbour et al., 1988; Tsuchiya et al., 1992; Sato et al., 1994; Shiseki et al., 1994). LOHs on Chromosome 17p affecting the p53 gene are also found frequently in human lung adenocarcinomas (from 28 to 70% of cases) and in other histotypes of lung tumors. Again, mutations in the remaining allele of the p53 gene have been observed in a number of cases (Yokota et al., 1987; Takahashi et al., 1991; Sato et al., 1994; Shiseki et al., 1994). 111. Genetic Linkage Studies

Inbred mice represent a good model system for the identification of the number and chromosomal localization of genetic loci predisposing lung tumor development. Indeed, linkage studies may be carried out by crossing two parental strains with large phenotype differences. The resulting F1 mice are then crossed together to obtain an F2 generation, characterized by the segregation of the phenotypic trait and the parental alleles at any genetic locus, including loci affecting the phenotype (lung tumor susceptibility). Alternatively, F1 mice may be back-crossed to one of the parental strains or to a third strain with a recessive or null phenotype, producing a back-cross or a test cross population, respectively. We crossed the A/J strain with the genetically resistant C3H/He strain. T h e resulting F2 population was treated with a single low dose of urethane, which induces many tumors in susceptible strains but very few, if any, in resistant strains (Dragani et al., 1991). The lung tumor susceptibility phenotype was evaluated quantitatively by using different parameters (Dragani et al., 1991). We typed 90 genetic markers, dispersed over the whole autosomes, and mapped a “major” locus associated with lung tumor development (Pasl)on the distal part of mouse Chromosome 6, near the Kras2 gene. (Fig. 1, Table 11). No other chromosomal region was linked to lung tumor susceptibility (Gariboldi et al., 1993). T h e Pasl locus explained u p to 45% of the variance in our cross and was supported by an LOD score >9. Sensitivity to urethane-induced lung tumorigenesis has previously been found to correlate with a Kras2 RFLP in AXB and BXA RI strains and in (C57BL/6 X A/J)F2 mice. However, this correlation was not confirmed in (C57BL/6 x A/J) x C57BL/6 back-cross mice, and no genetic markers other than Kras2 have been typed. The authors concluded that Kras2 may be one of the minor Pas genes (Ryan et al., 1987). I n our linkage study, the low genetic divergence and the consequent low degree of polymorphism between the parental strains make it diffi-

96

TOMMASO A. DRAGANI E T A L .

1

lOcM

-

Mtv23

-

Tpm3-rs3

hfet

-

CdSb

-

-

D6MitlO Rafl

-

Krm2

-

1)61ntl

I

v

NxV 1

I

9

I

I

7

5

D6Mitl3

3

LODscore

FIG. 1 . Genetic localization of the lung tumor susceptibility locus Pas1 to mouse Chromosome 6. T h e LOD score curves, above the threshold of 3, for the linkage of total lung tumor volume (NxV, dotted line) and lung tumor volume (V, solid line) to the genetic markers are shown on the left of the chromosome and are generated by the MAPMAKERIQTL computer package. T h e chromosomal region in which LOD scores 2 3 is shown with hatched lines. Distances between adjacent markers in centimorgans (cM) are based upon recombination fractions calculated by the MAPMAKER/EXP program (Haldane’s function),

cult to find genetic markers in all chromosomal regions. To improve the possibility of finding polymorphic markers, we repeated the experiment in an interspecific murine population that included the Mus spretus mice, a strain evolutionarily distant from laboratory mice that offers a great level of allelic polyniorphisms (Avner el al., 1988). Since the relative susceptibility to lung tumor development of M. spretus was unknown, instead of back-crossing the (A/J x M. spretus)Fl female mice to M. spretus, we chose to mate them with males of the C57BL/6J strain that were themselves resistant but that produced a susceptible F1 when crossed with susceptible strains (Bloom and Falconer, 1964; Malkinson, 1991). In this way, w e expected to obtain an analyzable sample, unless the M. spretus was susceptible, i.e., carried the same lung tumor susceptibility alleles of the A/J strain. We included in our analysis two groups of

97

GENETICS OF MURINE LUNG TUMORS

TABLE I1 MURINE LOCICONTAINING PUTATIVELUNGTUMOR SUSCEPTIBILITY~RESISTANCE GENES As DETECTED BY GENETIC LINKAGE STUDIES Variance expI aine d

Locus name

Chromosome

Cross

(W

Reference

~

6

Pas1

(A1J X C3HIHe)FZ (A/J X C57BL16)FZ (A/J X C57BL16) X C57BL16J (A/J X M. spctus) X C57BL16J

9

Pm4

(AJJ

11

Par1

(A/J X M. spctw) x C57BL16J

17 19

Pas2 Pas3

(A/J X C57BLI6)FZ (A/J X C57BL16)FZ (A1J X C57BL16) X C57BL16J

X

C57BL/6)F2

~ 4 0 % =60% =16%

Gariboldi ct al. (1993) Festing et al. (1994) Devereux ct al. ( 1994) (see note)

=4% 315% =7%

-2% =3%

Festing ct al. (1994) (see note) Festing et al. (1994) Festing ct al. (1994) Devereux ct al. (1994)

~~

No&. Data from our laboratory, manuscript in preparation.

(A/J x C57BL/6J)Fl (AB) and (C57BL/6J X M. spetw)Fl (BS) mice as positive and negative controls, respectively. T h e control mice represent the extreme possible genotypes of our test cross. As expected, the AB hybrids were susceptible, reaching 100% of tumor incidence in both male and female mice. The BS group did not develop any lung tumors, showing that this hybrid is genetically resistant to pulmonary tumorigenesis. Furthermore, because the C57BL/6J strain behaves recessively in the expression of the lung tumor susceptible phenotype in many hybrids tested for lung tumor incidence (Bloom and Falconer, 1964; Malkinson, 1991), our data indicated that the M.sfwetus itself is a resistant strain. Nearly 50% of (A/J X M. spretw) X C57BL/6J (ASB) test cross mice developed lung tumors, and lung tumor incidence was 80% in mice with the A/J allele at the Kras2 locus. This is compatible with the presence of one highly penetrant major susceptibility gene deriving from the A/J strain. Indeed, in the ASB test cross we confirmed the location of a Pas2 locus on the distal region of Chromosome 6. This locus accounted for 33.7% of the total variance observed in the population studied, with an LOD score of 8.9 that peaked on Kras2. However, when comparing the AB with the ASB mice, the values of the quantitative parameters associated with lung tumor susceptibility indicated the exis-

98

TOMMASO A. DRAGANI E T AL.

tence of one or more M . spretus alleles that strongly reduced the expressivity of the susceptible phenotype. Our observation adds to previously reported cases a new example of lung tumor resistance behaving as a dominant trait (Heston, 1942; Malkinson and Beer, 1983). We have obtained evidence for the existence on Chromosome 11 of a lung tumor resistance locus (Purl) derived from M . spretus (Table 11). This locus strongly decreased the expressivity of the Pus1 allele, but it did not affect lung tumor susceptibility in mice that do not carry the susceptible Pasl allele (Dragani et al., manuscript in preparation). Data from Festing et al. (1994) showed that at least four genes are associated with susceptibility to lung carcinogenesis in the (AIJ x C57BL/6)F2 cross. They confirmed the location near the Krm2 locus on Chromosome 6 of the “major” Pas1 locus, which accounted for 60% of the total variation in their cross. In addition, they found three “minor” loci associated with lung tumor development on Chromosomes 9, 17 (in the H 2 complex), and 19 (Table 11). They found no significant associations with 32 other loci located on all autosomes (Festing et al., 1994). Devereux et al. (1994) confirmed the mapping of Pasl and of the “minor” locus on Chromosome 19 (Table 11). Collectively, these results obtained with three different crosses from two laboratories confirm that the Pas1 locus plays a major role in lung tumor susceptibility. Genetic susceptibility to lung tumor development is not limited to the AIJ strain; some other strains also show high or intermediate susceptibility (Della Porta et al., 1967; Malkinson, 1989; Dragani et al., 1991). In particular, the SWR/J strain is as highly susceptible to lung tunior development as the A/J strain, although the two strains are phylogenetically distant (Atchley and Fitch, 1991). In fact, AIJ is an Mus musculus muscul w , whereas the SWRIJ is an M w mwculur domesticus (Nishioka, 1987; Kunieda and Toyoda, 1992). ‘Therefore, the genetic alterations leading to the high susceptibility to lung tumor development in the two strains could have originated independently. We are carrying out a genetic linkage study to map lung tumor susceptibility loci in the SWRiJ strain. For this experiment, we chose the BALB/c strain, an intermediate susceptible strain to lung carcinogenesis (Malkinson, 1989),as the other parental strain to obtain an F2 generation. T h e choice of the BALBlc strain was suggested by the fact that its intermediate susceptibility to lung carcinogenesis is dominant (Malkinson and Beer, 1983). Therefore, the genetic analysis of the F2 population would also allow us to find the chromosomal localization of the BALBlc dominant loci that partially suppress the SWRIJ loci predisposing to lung carcinogenesis. Our preliminary results on the (BALB/c X SWR/J)F2 population indicate that no linkage is present between lung tumor susceptibility and the Kras2 region of Chro-

GENETICS OF MURINE LUNG TUMORS

99

mosome 6.This result could be interpreted that both strains (SWR/J and BALB/c) contain the same Pas1 allele and, therefore, cannot be distinguished in the F2 cross (Manenti et al., manuscript in preparation). Several studies have indicated that the H2 complex, or genes close to H2 on Chromosome 17, affects lung tumorigenesis in mice (Miyashita and Moriwaki, 1987; Miyashita et al., 1989; Oomen et ad., 1991). Two reports, one on (A/J x C57BL/6)F2 mice and the other on two H2 congenic strains, confirmed the location in the H2 region of a putative “minor” lung tumor susceptibility locus (Festing et al., 1994; Fineman et al., 1994b). However, we have not found associations between loci close to H 2 and lung tumor susceptibility in both AC3F2 and ASB crosses (Gariboldi et al., 1993; data not shown). The discrepancy may be due to the different crosses and experimental schedules. IV. Transgenic Models of Lung Turnorigenesis So far, a number of transgenic lines of mice that develop lung tumors at a high incidence have been described. Contrary to the expectations of the researchers, three studies reported a high incidence of lung tumors in transgenic mice with the mutated HRAS gene under the control of different promoters lacking lung specificity, i.e., the immunoglobulin gene enhancer, the SV40 early gene promoter, the long terminal repeat of mouse mammary tumor virus (MMTV-LTR), and the albumin promoter (Suda et al., 1987; Tremblay et d.,1989; Sandgren et al., 1989). T h e genetic background of these mice varied, i.e., C57BL/6 x CD-1, (C3H x C57BL/6)F2 X BALB/c, and (C57BL/6 X SJL)Fl (Suda et al., 1987; Tremblay et al., 1989; Sandgren et ul., 1989). These strains are resistant or intermediately susceptible to lung carcinogenesis, but the transgenic lines derived from them developed lung tumors with high incidence and multiplicity and with early onset. Transgenic mice expressing the SV40 large T antigen in lung were produced. The SV40 large T antigen in one case was under the control of the human surfactant protein C (SP-C) gene promoter (Wikenheiser et al., 1992), and in the other case, it was under the control of the rabbit uteroglobin promoter (DeMayo et al., 1991). The rabbit uteroglobin gene is homologous to the 10-kDa Clara cell protein gene CClO (Wolf et ai., 1992). In both cases, the transgenic mice developed lung tumors at high incidence and early onset (3-12 months of age). The genetic background of these transgenic mice was the FVB strain, whose susceptibility to lung tumors is not known (DeMayo et al., 1991; Wikenheiser et al., 1992). Histologically, the lung tumors that developed in these transgenic mice invariably showed the common features typical of mouse lung tu-

100

TOMMASO A. DRACANI E T A L .

mors, i.e., papillary and solid patterns resembling the human adenocarcinoma histotype. A deletion mutant of p53 gene cloned from a Friend erythroleukemia cell line was used to generate transgenic mice in outbred CD-1 mice (Lavigueur et al., 1989). The construct carried a deletion of exon 2 and encoded a protein of 44 kDa, with a long half-life. The authors suggested that these mice develop lung tumors with relatively high inci, age at onset 11 months). dence (10 out of 112 mice, i.e., ~ 9 %mean Histological appearance was similar in all lung tumors, and they were classified as adenocarcinomas. However, the authors did not provide a control group of CD-1 mice, whose average incidence of spontaneous lung tumors is >lo% (range 0-4196) (Percy and Jones, 1971; Weisburger et al., 1978; Sher, 1982; Sher et al., 1982; Drew et al., 1983). The transgenic mice developed other tumor types, particularly osteosarcomas and lymphomas, with an overall tumor incidence of 20% (Lavigueur et al., 1989). However, transgenic mice with both alleles of the p53 tumor suppressor gene (frequently mutated in human lung cancer), which were knocked out by gene targeting, developed a variety of tumors but not lung tumors, which are rare in these mice (Donehower et al., 1992; Hursting et al., 1994). A transgenic mouse model has been established with tissue-specific expression of the LacZ gene. These mice contain a H2-KblLacZ fusion gene that shows lung-specific expression from the embryonic period through adulthood. Histochemical and immunocytochemical analyses indicated that type I1 pneumocytes were the only cell lineage showing LacZ expression (Hansbrough et al., 1993). This transgenic line may constitute a useful model for the study of the cell lineage of mouse lung tumors, as well as some biochemical and molecular aspects of gene expression in type I I pneumocytes. Transgenic models of lung tumorigenesis, as well as the available high incidence inbred strains (A, SWR), may be useful for the study of the pathogenesis of lung tumors and of potential chemotherapeutic and chemopreventative agents and as a bioassay system for chemical carcinogens. V. Candidate Lung Tumor Susceptibility Genes

‘The knowledge of the genetics of lung tumor susceptibility in mice is growing very quickly. Thus, in the near future we should be able to identi€y the chromosomal localization of most of the loci that affect lung carcinogenesis either positively or negatively. Such an analysis could

GENETICS OF MURINE LUNG TUMORS

101

provide information about the biological mechanism of susceptibility, suggesting possible candidate genes responsible for this phenomenon. Among the mapped loci affecting lung tumor susceptibility, we can speculate on a putative lung tumor susceptibility gene only for the Pas2 locus. Three kinds of indications would suggest that Kras2 may represent a candidate gene for the Pas2 locus: (i) Analysis of genetic linkage experiments performed in our laboratory on two distinct populations and in another laboratory on a third population have located a “major” genetic element affecting lung tumor susceptibility near the Kras2 locus (see Section 111). (ii) Kras2 is certainly involved in the pathogenesis of both spontaneous and chemically induced lung tumors (see Section II.B.l). The mutation rate observed in susceptible and resistant animals does not significantly correlate with lung tumor susceptibility (Stowers et al., 1987; Goodrow et al., 1990; Devereux et al., 1991, 1993). However, the reported specific activation of the Kras2 allele, derived from the sensitive parental strains in AC3F1 and C3AF1 hybrids, is intriguing and may suggest the existence of genetic elements affecting Krm2 mutations in the Kras2-susceptible or -resistant alleles (You et al., 1992a; Chen et al., 1994b). These differences may also account for the increased level of Kras2 transcript belonging to the sensitive allele in both normal and neoplastic lung tissues found in the same hybrids (You et al., 1992a; Chen et al., 1994b). (iii) Specific polymorphisms have been found in the second intron of the murine Kras2 that distinguish susceptible from resistant inbred and recombinant inbred strains (Ryan et al., 1987; Malkinson, 1991; Chen et al., 1994a). Indeed, resistant strains are characterized by the presence of a 37-base pair (bp) duplication [located at nucleotides 320 and 356 of the published sequence (You et al., 1992a)l and by distinctive bases (nt 288, A; nt 296, C; nt 494, T). However, sensitive strains lack the 37-bp direct repeat and differ from the Kras2resistant allele for the subsequent base changes G, C, and T in positions 288, 296, and 494, respectively (Chen et al., 1994a). Furthermore, the tandem repeat, typical of resistant strains was suggested as an important genetic element since gel retardation and DNAase I protection assays showed that it can bind nuclear proteins (Chen et al., 1994a). All of these observations raise the question of whether Kras.2 is the Pm2 locus or, alternatively, whether it represents a genetic marker for a closely linked, still unidentified gene. In their working hypothesis, Malkinson and You (1994) assign a principal role to the 37-bp repeat in lung tumor susceptibility. Its lack could lead to tumor development, affecting the likelihood of Kras2 mutability or enhancing the transcription-maturation of the mutated Kras2 allele

102

TOMMASO A. DRAGANI E T AL.

by several mechanisms. However, we found that the genomic structure of the second intron of the Kras2 allele in M. spretus has the specific features of a susceptible strain. It lacks the 37-bp direct repeat and has a C and an A at positions 288 and 296, respectively. Nevertheless, we found that M . spretus behaves like a resistant strain. T h e biological effect of a Pas1 allele derived from M. spetus differs from that of the A/J susceptible strain, as indicated by the linkage analysis. Moreover, among 13 urethane-induced tumors from (A/J x M. spretus)Fl,we found that in all eight samples with an activated Kras2, the mutation (codon 61) involved the A/J allele (Manenti et al., 1995). Therefore, the 37-bp repeat is not enough to explain either the difference in lung tumor susceptibility or the parental bias in Kras2 mutability between M. spretus and A/J alleles. In conclusion, the data available show that Kras2 is a genetic marker closely associated with the Pas1 locus. The identity of the putative gene for the Pasl locus remains undetermined. Kras2 continues to be a possible candidate gene since the genomic structures described may be part of the lung tumor susceptibility gene, yet separated from the element(s) controlling the genetic propensity to develop lung cancer. Alternatively, the biological function of Kras2 is related to the pathology but not to the genetics of lung tumors in mice, and the Pasl is a locus for a still unidentified gene. VI. Genetics of Lung Tumors, Human

Lung tumor is a relatively common type of cancer in humans, and familial clustering of cases is rare compared to colon and breast cancer, where both nonhereditary and familial cases are recognized. Exposure to environmental carcinogens, including tobacco smoking, represents the main risk factor for lung tumors (IARC, 1986). However, not all smokers develop lung cancer, and it is possible that other risk factors, including genetic factors, are implied in the pathogenesis of lung tumors in humans. Several studies have considered the possible role of genetic factors in human lung cancer risk, as reviewed in Law (1990) and Amos et al. (1992). These studies can be divided into familial studies and casecontrol studies. However, most of them have not taken into consideration the different histotypes of lung cancer, which may represent real differences in their pathogenesis and risk factors, including genetic ones. T h e risk of lung cancer in relatives of lung cancer patients has been

GENETICS OF MURINE LUNG TUMORS

103

examined in case-control studies by comparing first-degree relatives in case families with control families. Usually, the spouses of the cases were taken as controls. Tokuhata and Lillienfeld (1963) demonstrated a 2-2.5 excess risk of lung cancer in smoking relatives of cases compared with smoking relatives of controls. Lynch et al. (1986) reported a significant increase in cancers of all anatomic sites among the relatives of lung cancer probands. However, in the same relatives, they did not find evidence for increased lung tumor risk (Lynch et al., 1986). Ooi et al. (1986) studied first-degree relatives of lung cancer cases and their spouses. The c a s erelatives had a relative risk of 2.4 for lung cancer when compared with the control relatives, after adjusting for smoking and occupational exposure. Sellers et al. (1988) showed a relative risk of 2.5 for lung cancer among siblings of lung cancer cases when compared to siblings of the controls’ spouses. By segregation analyses of the same families studied previously (Ooi et al. 1986), Sellers et al. (1990) suggested that, after tobacco exposure, the pattern of lung cancer is best explained by Mendelian codominant inheritance of a single autosomal locus that influences the age at onset of lung cancer (Sellers et al., 1992). The results of their analysis indicated that (i) genetic predisposition to lung cancer is expressed only in the presence of tobacco smoke, and therefore, lung cancer is the result of a gene-environment interaction, and (ii) the influence of genetic factors in lung cancer pathogenesis is much greater than previously estimated, and most lung cancers may occur among gene carriers. However, they did not test genetic models that included polygenic inheritance of the characteristic. Rare cases of lung cancer clustering in twins and siblings have been reported (Brisman et al., 1967; Joishy et al., 1977; Paul et al., 1987; Biran et al., 1991). Three reports showed a common histotype of alveolar cell carcinoma (Joishy et al., 1977; Paul et al., 1987) or squamous cell carcinoma (Brisman et al., 1967) in affected cases, whereas another report showed different histotypes, although three-fourths of cases were NSCLCs (Biran et al., 1991). Two families with a high occurrence of lung tumors, as well as other cancers, were reported (Goffman et al., 1982). In one family, 5/10 siblings had lung tumors. In the second family, 4/8 siblings had lung cancers, which also occurred in 3/11 members older than 40 of the next generation. In both families, different histotypes of lung cancers, mostly NSCLCs, were observed. Lung cancer was associated with smoking habits in all cases (Goffman et al., 1982). A number of case-control studies have been reported on the possible association between lung cancer risk and particular haplotypes or phenotypes of genes coding for enzymes involved with drug metabolism,

104

TOMMASO A. DRAGANI E T AIL

including carcinogen metabolism. These studies are based on the fact that cigarette smoking is associated with an increased risk for different histological types of lung tumors, including lung adenocarcinomas JIARC, 1986). Cigarettes contain a number of different chemical carcinogens that must be activated by endogenous enzymes to form reactive chemical species capable of covalently binding to DNA to cause mutations. Indeed, mutations at the Krm2 gene have been identified in a high percentage of lung adenocarcinomas, and the presence of KRAS2 mutations has been associated with cigarette smoking (Slebos et al., 1990; Rodenhuis and Slebos, 1992; Husgafvel-Pursiainen et al., 1993). Since the metabolism of chemical carcinogens involves a variety of phase I and phase I1 enzymes, genotypic differences at the loci coding for these enzymes may lead to differences in the endogenous activation of chemical carcinogens and, consequently, to differences in lung cancer risk. A possible association between genetic predisposition to lung cancer and the inheritance of specific alleles or phenotypes at P450 (CYPlAl, CYP2D6, C’t’P2E l), and glutathione S-transferase (GSTM 1) loci has been reported (Caporaso et al., 1990; Kawajiri et al., 1990; Uematsu et al., 1991; Ketterer et al., 1992; Hirvonen et al., 1993b; Kihara et al., 1993). However, some other studies have failed to confirm these associations (Sugimura et al., 1990; Hirvonen et al., 1992, 1993a; Kato et al., 1992, 1994; Brockmoller et al., 1993; Anttila et al., 1994). The increased risk reported in the positive studies varied from 2 to 3. Comparative mapping indicated that the human 1 2 ~ 1 chromosomal 2 region (around KRAS2) is homologous to the mouse region containing Pasl (Gariboldi at al., 1993; Copeland et al., 1993; MGD, 1994). Therefore, w e have designed a case-controt study to test for possible associations between RFLPs at the KRAS2 locus and the risk of adenocarcinoma, the human tumor histotype that closely resembles the mouse lung tumor histotype. We have studied 120 Italian cases and 120 controls matched by the same geographical origin, and we have typed two KRAS2 RFLPs. There was no association between the risk of lung adenocarcinoma and particular haplotypes at the KRAS2 locus (Manenti et nl., manuscript in preparation). These negative results, however, do not exclude the possible existence in this chromosomal region of a lung tumor susceptibility locus. In fact, it is possible that the putative locus is not in linkage disequilibrium with KRAS2 haplotypes, and therefore, it cannot be detected by association (case-control) studies. Linkage studies in lung cancer pedigrees and in affected sibling pairs may be more appropriate to test for the possible location of a hunian Lung cancer susceptibility gene near KRAS2.

GENETICS OF MURINE LUNG TUMORS

105

VII. Conclusions and Perspectives Several studies have indicated the possible important role of genetic factors in human lung cancer risk. However, familial lung cancer is rare, and the lack of familial clusters of lung cancer patients indicates the low penetrance or the multigenic nature of inherited predisposition to lung cancer. Therefore, the identification and cloning of single genetic elements affecting predisposition to lung cancer cannot be accomplished in humans, because we cannot identify the genetically susceptible individuals without the familial clustering of cases. T h e murine strains predisposed to lung tumor development may provide a unique experimental system for the analysis of the genetics of these tumors. I n fact, after inbreeding, these strains carry, at the homozygous status, the alleles predisposing lung tumorigenesis. These alleles can only be selected and identified by the inbreeding process, but cannot be recognized in the outbred wild population. Indeed, due to the genetic heterogeneity of the individuals in an outbred population, it cannot be assessed whether an animal develops a cancer simply by chance or because it is genetically predisposed. Murine lung tumors represent the experimental counterpart of human lung adenocarcinomas, since these two tumor types share common patterns of histology, mutations at the Kras2 gene, specific alterations o r maintenance of expression of different genes, and LOHs at mousehuman homologous chromosomal regions. Although a complete overlap in the characteristics of mouse lung tumors and human lung adenocarcinomas does not occur, the common features outweigh the differences. A “major” genetic locus affecting susceptibility to the development of lung tumors (Pasl)has been mapped on the mouse genome by genetic linkage analysis experiments. Other “minor” susceptibility loci and loci affecting dominant resistance to lung carcinogenesis are presently being mapped. Finally, the results of genetic linkage studies could provide a clear picture of the number and chromosomal location of loci affecting lung tumorigenesis in the experimental system. This will be the first step toward the cloning of lung tumor susceptibility and resistance genes. Although the positional cloning of the Pasl locus affecting lung tumor development is difficult, due to the relatively large regions of linkage (5-10 cM, which may contain hundreds of genes), this goal is possible. Indeed, the obese (ob) gene has been cloned on the basis of results of genetic linkage experiments (Zhang et al., 1994). New developments in the methodologies for the positional cloning of genes may allow us in the near future to clone the Pasl gene and other “minor” loci affecting lung tumor development.

106

TOMMASO A. DRAGANI E T A L ,

Once the genes responsible for susceptibility to lung tumorigenesis in mice are cloned, it is relatively easy to find the human homologues and to look for the presence of mutations and/or allelic variations in cancer patients and in the general population. T h e identification of the subject at genetic risk for lung cancer may be useful to implement cancer prevention strategies in people at genetic risk. ACKNOWLEDGMENTS This work \\as supported in part by grants from the Italian CNR (PF "ACRO") and Associazione Italiana Ricerca Cancro.

REFERENCES

Amos, C . I., Caporaso, N . E., and Weston, A. (1992) Cancer Epzdem. Biomar. 1, 505-513. Andervont, H . B. (1937). Publ. Health Rep. (Washington D.C.) 52, 212-221. Andervont, H. B. (1940).J. iWl. Cancer Imt. 1, 135-145. Anttila, S., Hirvonen, A., Husgafvel-Pursiainen, K., Karjalainen, A,. Nurminen, T., and Vainio, H. (1994). Carcinogen& 15, 1133-1 135. Atchley, W.R., a n d Fitch, W. M. (1991). Science 254, 554-558. Avner, P., Amar, L., Dandolo, L., and Guenet, J. L. (1988). Trends Genet. 4, 18-23. Barbacid, M. (1987). Annu. Rev. Biochetn. 56, 779-827. Beer, D. G., and Malkinson, A. M. (1985).J. Natt. Cancer Iwt. 75, 963-969. Belinsky, S. A., Devereux, T. R., Maronpot, R. R., Stoner, G. D., and Anderson, M. W. (1 989). Canrtr Reg. 49, 5305-53 1 1 . Belinsky S. A . , Devereux, T. R., Foley, J. F., Maronpot, R. R., and Anderson, M. W. (1992a). Canrer RPS. 52, 3164-3173. Belinsky, S. A.. Devereux, T. R., Fole); J. F., Maronpot, R. R., and Anderson, M. W. (1992b). Cancer Res. 52, 3164-3173. Belinsky, S. A., Stefanski, S. A,, and Anderson, M. W. (1993). Cancer Reg. 53, 410-416. Biran, H., Goldstein, J., and Cohen, Y. (1991). Lung Cancer 7, 345-353. Bittner, J. J. (1938). Publ. Health Rep. (Warhzngton DC) 53, 2197-2202. Bittner, J. J. (1939). Puhl. Health Rep. (Washington DC) 54, 380-392. Bloom, J. L., a n d Falconer, D. S. (1964). J. h'atl. Cancer Imt. 33, 607-618. Brisman, R., Baker. R. R., Elkins, R., and Hartmann, W. H. (1967). Cancer20,2048-2053. Brockmoller, J., Kerb, R., Drakoulis, N., Nitz, M., and Roots, I. (1993). Cancer Res. 53, 1004-1011. Broers, J. L. V., Jensen, S. M.,Travis, W. D., Pass, H., Whitsett, J. A., Singh, G., Katyal, S. L., Gazdar, A. F., Minna, J. D., and Linnoila, R. I. (1992). Lab. Invest. 66, 337-346. Buchhagen, D. L. (1991). Biochim. Biophys. Acta 1072, 159-176. (:aims, P., Mao, L., Merlo, A , , Lee, D. J., Schwab, D., Eby, Y., Tokino. K., van d e r Riet, P., Blaugrund, J. E., and Sidransky, D. (1994). Science 265,415-416. Candrian, U., You, M., Goodrow, T., Maronpot, R. R., Reynolds, S. H., and Anderson, M. W. (1991). CanrerRes. 51, 1148-1153. (:aporaso, N. E., Tucker, M. A,, Hoover, R. N., Hayes, R. B., Pickle, L. W., Issaq, H. J., Muschik. G. M., Green-Gallo, L., Buivys, D., Aisner, S., Resau, J. H., Trump, B. F., Tollerud, D., Weston, A., and Harris, C. C. (l99O).J. Natl. Cancerlnst. 82, 1264-1271.

GENETICS OF MURINE LUNG TUMORS

107

Chen, B., Liu, L., Castonguay, A., Maronpot, R. R., Anderson, M. W., and You, M. (1993). Carcinogenesis 14, 1603-1608. Chen, B., Johanson, L., Wiest, J. S., Anderson, M. W., and You, M. (1994a).Proc. Natl. Acud. Sci. USA 91, 1589-1593. Chen, B., You, L., Wang, Y., Stoner, G. D., and You, M. (1994b). Carcinogenesis 15, 20312035. Chiba, I., Takahashi, T., Nau, M., DAmico, D., Curiel, D., Mitsudomi, T. D. B., Carbone, D., Piantadosi, S., Koga, H., Reissmann, P., Slamon, D., Holmes, S., and Minna, J. (1990). Oncogene 5, 1603-1610. Copeland, N. G., Jenkins, N. A., Gilbert, D. J., Eppig, J. T., Maltais, L. J., Miller, J. C., Dietrich, W. F., Weaver, A., Lincoln, S. E., Steen, R. G., Stein, L. D., Nadeau, J. H., and Lander, E. S. (1993). Science 262, 57-66. Della Porta, G., Capitano, J., Parmi, L., and Colnaghi, M. I. (1967). Tumori 53, 81-102. DeMayo, F. J., Finegold, M. J., Hansen, T. N., Stanley, L. A,, Smith, B., and Bullock, D. W. (1991).Am. J. Physiol. 261, L70-L76. Der, C. J., Finkel, T., and Cooper, G. M. (1986). Cell 44, 167-176. Devereux, T. R., Anderson, M. W., and Belinsky, S. A. (1991). Carcznogenesis 12,299-303. Devereux, T. R., Belinsky, S. A., Maronpot, R. R., White, C. M., Hegi, M. E., Patel, A. C., Foley, J. F., Greenwell, A., and Anderson, M. W. (1993). Mol. Carcinogen. 8, 177-185. Devereux, T. R., Wiseman, R. W., Kaplan, N., Garren, S., Foley, J. F., White, C. M., Anna, C., Watson, M. A., Patel, A,, Jarchow, S., Maronpot, R. R., and Anderson, M. W. (1994). Mamm. Genome 5,749-755. Donehower, L. A,, Harvey, M., Slagle, B. L., McArthur, M. J., Montgomery, C. A., Butel, J. S., and Bradley, A. (1992).Nature 356, 215-221. Dragani, T. A., Manenti, G., and Della Porta, G. (1991). Cancer Res. 51, 6299-6303. Drew, R. T., Boorman, G. A., Haseman, J. K., McConnell, E. E., Busey, W. M., and Moore, J. A. (1983). Toxicol. Appl. Phannacol. 68, 120-130. Festing, M. F. W., Yang, A,, and Malkinson, A. M. (1994). Genet. Res., in press. Fijneman, R. A., Ophoff, R. A., Hart, A. A. M., and Demant, P. (1994a).Oncogene 9,14171421. Fijneman, R. J. A,, Oomen, L. C. J. M., Snoek, M., and Demant, P. (1994b).Immunogenetics, 41, 106-109. Gariboldi, M., Manenti, G., Canzian, F., Falvella, F. S., Radice, M. T., Pierotti, M. A., Della Porta, G., Binelli, G., and Dragani, T. A. (1993). Nature Genet. 3, 132-136. Gazdar, A. F. (1994). Lung Cancer 11, 16-17. Gazdar, A. F., Linnoila, R. I., Kurita, Y., Oie, H. K., Mulshine, J. L., Clark, J. C., and Whitsett, J. A. (1990). Cancer Res. 50, 5481-5487. Goffman, T. E., Hassinger, D. D., and Mulvihill, J. J. (1982).JAMA247, 1020-1023. Goodrow, T. L., Reynolds, S., Maronpot, R., and Anderson, M. (1990). Cancer Res. 50, 48 18-4823. Goodrow, T. L., Storer, R. D., Leander, K. R., Prahalada, S. R., van Zwieten, M. J., and Bradley, M. 0. (1992).Mol. Carcinogen. 5, 9-15. Gosney, J. R., Field, J. K., Gosney, M. A., Lye, M. D. W., Spandidos, D. A., and Butt, S. A. (1990). Anticancer Res. 10, 623-628. Grady, H. G., and Stewart, H. L. (1940). Am.J. Pathol. 16,417-432. Hansbrough, J. R., Fine, S. M., and Gordon, J. I. (1993).J. Biol. Chem. 268, 9762-9770. Harbour, J. W., Lai, S.-L., Whang-Peng, J., Gazdar, A. D., Minna, J. D., and Kaye, F. J. (1988). Science 241, 353-357. H'egi, M. E., Soderkvist, P., Foley, J., Schoonhoven, R., Swenberg,J. A., Kari, F., Maronpot, R., Anderson, M. W., and Wiseman, R. W. (1993). Carcinogenesis 14, 803-810.

108

T O M M A S O A . D R A G A N I ET A L .

Herzog, C. R., Schut, H. ’4. J., Maronpot, R. R., and You, M. (1993). Mol. Carcinogen. 8, 202-207. Herzog. C. R., Wiseman, R. W., and You, M. (1994). Cancer Res. 54, 4007-4010. Heston, W. E. (1940).J . Null. Cancer Inst. 1, 105-1 1 1 . Heston, W. E. (1942).J . Natl. Cancer Inst. 3, 69-78. Hirvonen, A , , Husgafvel-Pursiainen, K., Karjalainen, A., Anttila, S., and Vainio, H. (1992). Cancer Eprdemtol., B i m r k e r s Preuention 1, 485-489. Hirvonen, A., Husgafvel-Pursiainen, K., Anttila, S . , Karjalainen, A,, and Vainio, H. (1993a). Carcznogenesu 14, 85-88. Hirvonen, A., Husgafvel-Pursiainen, K., Anttila, S., and Vainio, H. (1993b). Carcinogenesis 14, 1479-1481. Horio, Y . , Takahashi, T., Kuroishi, T., Hibi, K., Suyama, M., Niimi, T., Shimokata, K., Yamakawa, K., Nakamura, Y . , and Ueda, R. (1993). Cancer Res. 53, 1-4. Hursting, S. D., Perkins, S. N., and Phang, J. M. (1994).Proc. Natl. A d . Sci. USA 91,70367040. Husgafvel-Pursiainen, K., Hackman, P., Rindanpaa, M., Anttila, S., Karjalainen, A., and Partanen, 1..(1993). Int. /. Cancer 59, 250-256. Hussussian, C. J., Struewing, J. P., Goldstein, A. M., Higgins, P. A. T., Ally, D. S., Sheahan, M. D., Clark, W. H . , Tucker, M . A,, and Dracopoli, N. C. (1994)..Vature Genet. 8, 15-21. IARC (1986). “Monographs o n the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Tobacco Smoking,” Vol. 38, pp. 83-126. IARC, Lyon, France. Joishy, S. K., Cooper, R. A., and Rowlet; P. T. (1977). Ann. Intern. Med. 87, 447-450. Kamb, A., Shattuck-Eidens, D., Eeles, R., Liu, Q.,Gruis, N. A , , Ding, W., Hussey, C., Tran, T., Miki, Y., Weaver-Feldhaus,J., McClure, M., Aitken, J. F., Anderson, D. E., Bergman, W., Frants, R., Goldgar, D. E., Green, A,, MacLennan, R., Martin, N. G., Meyer, L. I., Youl, P., Zone, J. J., Skolnick, M. H., and Cannon-Albright, L. A. (1994).Nature Genet.8, 22-26. Kato, S . . Shields, P. G., Caporaso, N . E., Hoover, R. N., Trump, B. F., Sugimura, H., Weston, A., and Harris, C. C. (1992). Cancer Res. 52, 6712-6715. Kato, S., Shields, P. G., Caporaso, N. E., Sugimura, H., Trivers, G. E., Tucker, M. A,, Trump, B. F., Weston, A., and Harris, C. C. (1994). CancerEprdemiol., Biomarkers Preventzon 3, 51.5-318. Kawajiri, K., Kakachi, K., Imai, K., Yoshii. A., Shinoda, N., and Watanabe, J. (1990). FEBS 263, 131-133. Kern, J. A., Slebos, R. J. C . , Top, B., Rodenhuis, S., Lager, D., Robinson, R. A., Weiner, D., and Schwartz, D. A . (1994).J . Clin. Invest. 93, 5 16-520. Ketterer, B., Harris, J. M.,Talaska, G., Meyer, D. J.. Penible, S.E., Taylor, J. B., Lang, N. P., and Kadlubar, F. F. (1992). Envtron. Healfh Perspect. 98, 87-94. Kihara, M., Kihara, M.,Noda, K., and Okamoto, N. (1993) Cancer Lett. 71, 151-155. Kimura. I. (1971). Acta Pathol.Jpn. 21, 13-56. Kishirnoto, Y., hfurakami, Y . , Shiraishi, M., Hayashi, K., and Sekiya, T. (1992). Cancer Res. 52,4799-4804. Kobayashi, T., Tsuda, H., Noguchi, M., Hirohashi, S., Shirnosato, Y., Goya, T., and Hayata, Y. (1990). Cancer 66, 289-294. Kunieda, T., and Toyoda, Y. (1992). Genomzcs 13, 236-237. Kuroki, Y., and Voelker, D. R. (1994).J. Biot. Chem. 269, 25943-25946. Lavigueur. A., Maltby, V., Mock, D., Rossant, J., Pawson, T., and Bernstein, A. (1989). Mol. CPU.Biol. 9, 3982-3991. Law, M. R. (1990). Br..J. Cancer 61, 195-206.

GENETICS OF MURINE LUNG TUMORS

109

Leff, S. E., Brannan, C. I., Reed, M. L., Ozcelik, T., Franke, U.,Copeland, N. G., and Jenkins, N. A. (1992).Naiure Genet. 4, 259-264. Levinson, B., Vulpe, C., Elder, B.,Martin, C., Verley, F., Packman, S., and Gitschier, J. (1994).Naiure Genet. 6, 369-373. Li, E. E., Heflich, R. H., Bucci, T. J., Manjanatha, M. G., Blaydes, B. S., and Delclos, K. B. 15, 1377-1385. (1994a).CarcinogLi, Z.-H., Zheng, J., Weiss, L. M., and Shibata, D. (1994b).A m . ] . Pafhol. 144, 303-309. Li Lung, M., Wong, M., Kit Lam, W., Shing Lau, K., Kwan, S., Fu,K., Cheung, H., and Wei Yew, W. (1992).Cancer 70, 760-763. Linnoila, R. I., Mulshine, J. L., Steinberg, S. M., and Gazdar, A. F. (1992).J. Nail. Cancer Imt. MOTIOP. 13,61-66. Loeb, L. A., Ernster, V. L., Warner, K. E., Abbotts, J., and Laszlo, J. (1984).Cancer Res. 44,

5940-5958.

Lukeis, R., Irving, L., Garson, M., and Hasthorpe, S. (1990).Genes Chrom. Cancer4, 116-

124. Lynch, C . J. (1926).J. Exp. Med. 43,339-355. Lynch, H. T., Kimberling, W. J., Markvicka, S. E., Biscone, K. A., Lynch, J. F., Whorton, E., and Mailliard, J. (1986).Cancer 57, 1640-1646. Lyon, M. F.,Peters, J., Glenister, P. H., Ball, S., and Wright, E. (1990).Proc. Nail. Acad. Sci. USA 87,2433-2437. Malkinson, A. M. (1989).Taxicohgy 54,241-271. Malkinson, A. M. (1991).Enuirmr. f f e d f hPmpecf. 93, 149-159. Malkinson, A. M. (1992).Cancer Rex 54,2670s-2676s. Malkinson, A. M., and Beer, D. S. (1983).J. N d . Cancer Inrt. 70,931-936. Malkinson, A. M., and You, M. (1994).Mol. Carcinogen. 10, 61-65. Malkinson, A. M., Nesbitt, M. N.,and Skamene. E. (1985).J.Nad. CancerInrt. 75,971-974. Manam, S . , Storer, R.D., Prahalada, S., Leander, K.R., Kraynak, A. R., Hammermeister, C. L., Joslyn, D. J., Ledwith, B.J., van Zwieten, M. J., Bradley, M. O., and Nichols, W. W. (1992).Mol. Carcinogen. 6, 68-75. Manenti, G., Falvella, F. S., Gariboldi, M., Dragani, T. A,, and Pierotti, M. A. (1995). Genomics, in press. Maronpot, R. R., Shimkin, M.B., Witxhi, H. P.,Smith, L. H., and Cline, J. M.. (1986).J. Nail. C a n c e r h i . 76, 1101-1112. Mass, M. J., Jeffers, A. J., Ross, J. A., Nelson, G., Galati, A. J., Stoner, G. D., and Nesnow, S. (1993).Mol. Carcinogen. 8, 186-192. Matzinger, S. A., Gunning, W.T., You, M., and Castonguay, A. (1994).M o l . Carcinogen. 11,

42-48. Mead, L. J., Gillespie, M. T., Irving, L. B., and Campbell, L. J. (1994).CancerRes. 54,2307-

2309.

Merlo, A., Gabrielson, E.,Askin, F., and Sidransky, D. (1994).Cancer Res. 54, 640-642. MGD, Mouse Genome Database (1994).Release 1.1: T h e J a c k i n Laboratory, Bar Harbor, ME. Miller, C. W., Simon, K.,Aslo, A., Kok, K., Yokota, J., Buys, C. H., Terada, M., and Koeffler, H. P. (1992).Cancer Re.54, 1695-1698. Minna, J. D. (1993).Chcsl 103, 4458-456s. Mitsudomi, T., Steinberg, S. M., Oie, H. K., Mulshine, J. L., Phelps, R.,Viallet, J., Pass, H., Minna, J. D., and Gazdar, A. F. (1991).Cancer Ra.51,4999-5002. Miyashita, N., and Moriwaki, K. (1987).Jpn. J. Cancer k s . 78,494-498. Miyashita, N., Moriwaki, K., and Migita, S. (1989).Immunogeneiics 49, 14-18.

110

TOiMMASO A. DRAGANI E7’ A L .

Nishioka, Y. (1987). Genet. Res. 50, 69-72. Nuzum, E. O., Malkinson, A. M., and Beer, D. G. (1990). Mol. Carcinogen. 3, 287-295. Ohmori, H., Abe, T., Hirano, H., Murakami, T., Katoh, T., Gotoh, S., Kido, M., Kuroiwa, A., Nomura, T., and Higashi, K. (1992). Carcinogenesis 13, 851-855. Okamoto, A., Hussain, S. P., Hagiwara, K., Spillare, E. A , , Rusin, M. R., Demetrick, D. J., Serrano, M.. Hannon, G. J., Shiseki, M., Zariwala, M., Xiong, Y., Beach, D. H., Yokota, J., a n d Harris, C. C. (1995). Cancer Res. 55, 1448-1451. Olopade, 0. I., Buchhagen, D. L., Malik, K., Sherman, J.. Nobori, T., Bader, S., Nau, M. M., Gazdar. A. F., Minna, J. D., and Diaz, M. 0. (1993). Cancer Res. 53,2410-2415. Ooi, MI. L., Elston, R. C., Chen, V. W., Bailey-Wilson, J. E., a n d Rothschild, H. (1986).J. Natl. Cancer imt. 76, 2 f 7-222. Oomen, L. C. J. M.,van d e r Valk, M. A,, and Demant, P. (1991). Ex@ Lung Res. 17,283304. Paul, S. M., Bacharach, B., and Goepp, C. (1987).J . Surg. Oncol. 36, 249-252. Percy, D. H., and Jones, A. M. (1971).J. Natl. CancerInst. 46, 1045-1065. Pereira, M . A., and Stoner, G. D. (1985). Fund. Appl. Toxzcol. 5, 688-699. Quinlan, D. C . , Davidson, A. G., Summers, C. L., Warden, H. E.. and Doshi, H . M. (1992). Cancer Res. 52, 4828-483 I. Re, F. C., Manenti, G., Borrello, M. G., Colombo, M. P., Fisher,J. H., Pierotti, M. A., Della Porta, C . , and Dragani. T. A. (1992). Mol. Carcinogen. 5 , 155-160. Kehm, S . , Ward, J. M., Have-Ophroek, A. A. W. T., Anderson, L. M., Singh, G., Katyal, S., and Rice, J. M. (1988). Cancer Res. 48, 148-160. Reissmann, P. T., Koga, H., Takahashi, R., Figlin, R. A,, Holmes, E. C., Piantddosi, S., Cordon-Cardo, C., Slamon, D. J., and Lung Cancer Study Group (1993). Oncogene 8, 1913-1919. Revnolds, S. H., Anna, C. K., Brown, K. C., Wiest, J. S., Beattie, E. J., Pero, R. W., Iglehart, 1. D., and Anderson, M. W. (1991).Proc. Natl. Acad. Sci. LISA 88, 1085-1089. Rodenhuis, S . , and Slebos, R. J. C. (1992). Cancer Res. 52, 2665s-2669s. Rodenhuis, S . , van d e Wetering, M. L., Moot, W. J . , Evers, S. G., van Zandwijk, N., and Bos, J. L. (1987). XPUJEr~gl.J . Med. 317, 929-935. Rosai, J. (1989). in “Ackerman’s Surgical fathology,” 7th ed., pp. 264-343. C. V. Mosby Company, St. Louis. Rosell, R.. Li, S., Skacel, Z., Mate, J. L., Maestre, J., Canela, M., Tolosa, E., Armengol, P., Barnadas, A,, and Ariza, A. (1993). Oncogene 8, 2407-2412. Ryan, J . , Barker. P. E., Neshitt, hl. N., and Ruddle, F. H. (1987).J. Null. Cancer inst. 79, 135 1-1357. Ryberg, D., Kure, E., Lystad, S., Skaug, V., Stangeland, L., Mercy, I., Borresen, A.-L., and Haugen, A. (1994). Cancer Res. 54, 15.51-1555. Sandgren, E. P., Quaife, C. J., Pinkert, C. A,, Palmiter, R. D., and Brinster, R. L. (1989). Oncogeiae 4, 715-724. Sato, S.. Nakdmura, Y., and Tsuchiya, E. (1994). Cancer Res. 54, 56.52-5655. Sellers, T. A., Elston, R. C., Stewart, C., and Rothschild, H . (1988). Genet. Epzderniol. 5,381392. Sellers, T. A . , Bailey-Wilson, J. E., Elston, R. C.,Wilson, A. E., and Rothschild, H. (199O).J, Natl. Cancer In% 82, 1272- 1279. Sellers, T. A . , Potter, J. D., Beiley-Wilson, J. E., Rich, S. S., Rothschild, H., and Eltson, R. C. (1992). Cancer Res. 52, 2694s-2697s. Sher, S. P. (1982). Crit. Rat Toxzcol. 10, 49-79. Sher, S . P., Jensen, R. D., and Bokelnian, D. L. (1982). Tuxicol. Lett. 11, 103-110.

GENETICS OF MURINE LUNG TUMORS

111

Shijubo, N., Tsutahara, S., Hirasawa, M., Takahashi, H., Honda, Y., Suzuki, A., Kuroki, Y., and Akino, T. (1992). Cancer 69, 2905-2909. Sh.imkin, M. B., and Stoner, G. D. (1975).Adv. Cancer Res. 21, 1-58. Shliseki, M., Kohno, T., Nishikawa, R., Sameshima, Y., Mizoguchi, H., and Yokota, J. (1994). Cancer Res. 54, 5643-5648. Singh, G., Singh, J., Katyal, S. L., Brown, W. E., Kramps, J. A., Paradis, I. L., Dauber, J. H., Macpherson, T. A., and Squeglia, N. (1988).J. Histochem. C y t o c h . 36, 73-80. Slebos, R. J. C., Kibbelaar, R. E., Dalesio, O., Kooistra, A., Stam, J., Neifier, C. J. L. M., Wagenaar, S. S., Van der Schueren, R., van Zandwijk, N., Mooi, W. J., Bos, J. L., and Rodenhuis, S. (1990). New Engl. J.Med. 323, 561-565. Slebos, R. J. C., Hruban, R. H., Dalesio, O., Mooi, W. J., Offerhaus, G . J. A., and Rodenhuis, S. (1991).J. Natl. Cancer Inst. 83, 1024-1027. Smith, M. J., Rousculp, M. D., Goldsmith, K. T., Curiel, D. T., and Carver, R. I. (1994). Human Gene Ther. 5,29-35. Sozzi, G., Miozzo, M., Donghi, R., Pilotti, S., Cariani, C. T., Pastorino, U., Della Porta, G., and Pierotti, M. A. (1992). Cancer Res. 52, 6079-6082. Stewart, H. L., Dunn, T. B., Snell, K. C., and Deringer, M. K. (1979) “Pathology of Tumours in Laboratory Animals. Tumours of the Mouse” (V. S. Turusov, Ed.), pp. 25 1287. IARC Scientific Publications No. 23. IARC, Lyon, France. Stowers, S. J., Glover, P. L., Reynolds, S. H., Boone, L. R., Maronpot, R. R., and Anderson, M. W. (1987). Cancer Res. 47,3212-3219. Strong, L. C. (1936).J. Hered. 27, 21-24. Su, L.-K., Kinzler, K. W., Vogelstein, B., Preisinger, A. C., Moser, A. R., Luongo, C., Could, K. A., and Dove, W. (1992). Science 256,668-670. Suda, Y., Aizawa, S., Hirai, S.-i., Inoue, T., Furuta, Y., Suzuki, M., Hirohashi, S., and Ikawa, Y. (1987). EMBO J. 6,4055-4065. Sugirnura, H., Caporaso, N. E., Shaw, G. L., Modali, R. V., Gonzales, F. J., Hoover, R. N., Resau, J. H., Trump, B. F., Weston, A., and Harris, C. C. (1990). Carcinogenesis 11, 1527-1530. Sugio, K., Ishida, T., Yokoyama, H., Inoue, K., Sugimachi, K., and Sasazuki, T. (1992). Cancer Res. 52, 2903-2906. Takahashi, T., Suzuki, H., Hida, T., Sekido, Y., Ariyoshi, Y., and Ueda, R. (1991). Oncogene 6, 1775-1778. Xissabehji, M., Read, A. P., Newton, V. E., Patton, M., Gruss, P., Harris, R., and Strachan, T. (1993). Nature Genet. 3, 26-30. Testa, J. R., Siegfried, J. M., Liu, Z., Hunt, J. D., Feder, M. M., Litwin, S., Zhou, J.-y., Taguchi, T., and Keller, S. M. (1994). Genes Chrum. Cancer 11, 178-194. Thaete, L. G., Gunning, W. T., Stoner, G. D., and Malkinson, A. M. (1987).J. Natl. Cancer Inst. 78,743-749. Thaete, L. G., Nesbitt, M. N., and Malkinson, A. M. (1991). Cancer Lett. 61, 15-20. Tokuhata, G. K., and Lillienfeld, A. M. (1963).J.Natl. CancerInst. 30, 289-312. Tremblay, P. J., Pothier, F., Hoang, T., Tremblay, G., Brownstein, S., Liszauer, A,, and Jolicoeur, P. (1989). Mul. Cell. Biol. 9, 854-859. Tsuchiya, E., Nakamura, Y., Weng, S.-Y., Nagakawa, K., Tsuchiya, S., Sugano, H., and Kitagawa, T. (1992). Cancer Res. 52, 2478-248 1. Uematsu, F., Kikichi, H., Motorniya, M., Abe, T., Sagami, I., Ohmachi, T., Wakui, A,, Kanamaru, R., and Watanabe, M. (1991).Jpn.J. Cancer Res. 82, 254-256. Vogelstein, B., and Kinzler, K. W. (1992). Cell 70, 523-526. b l m , M., Drings, P., Wodrich, W., and van Kaick, G. (1993).Clin. Exp. Metastasis 11,325-329.

112

TOMMASO A . DRAGANI E T A L .

Wang, Y.. Stoner, G., and You, M. (1993). Cancer Res. 53, 1620-1624. Wang-Peng, J., Knutsen, T., Gazdar, A,, Steinberg, S. M., Oie, H., Linnoila, I., Mulshine, J., Nau, M., and Minna, J. D. (1991). Genes Chrom. Cancer 3, 168-188. Ward, J. M., Singh, G., Katyal, S. L., Anderson, L. M., and Kovatch, R. M. (1985). Am. J. Pathol. 118, 493-499. Weisburger, E. K., Russfield, A. B., Hornburger, F., Weisburger, J. H., Boger, E., Van Dongen, C. G., and Chu. K. C. (1Y78).J. Environ. Pathol. Toxicol. 2, 325-356. Westra, W.H., Slebos, R.J. C., Offerhaus, G. J. A,, Goodman, S. N., Evers, S. G., Kensler, T. W., Askin, F. B., Rodenhuis, S., and Hruban, R. H. (1993). Cancer 72, 432-438. Wikenheiser, K . A., Clark, J. C., Linnoila, R. I., Stahlman, M. T.,and Whitsett, J. A. (1992). Cancer Res. 52, 5342-5352. Wisenian, R. W., Cochran, C., Dietrich, W., Lander, E. S., and Soderkvist, P. (1994). Proc. Natl. Acad. Sci. USA 91, 3759-3763. Wodrich, W., and Volm, M. (1993). Carcinogenesis 14, 1121-1124. Wolf, M., Klug, J., Hackenberg, R., Gessler, M., Grzeschik, K. H., Beato, M., and Suske, G. (1992). Hum. M o 2 . Genet. 1, 371-378. Xu, H.-J., Hu. S.-X., Cagle, P. T., Moore, G . E., and Benedict, W. F. (1991). CancerRes. 51, 2735-2739. Xu, H.-J., Quinlan, D. C., Davidson, A. G., Hu, S.-X., Summers, C. L., Li, J., and Benedict, W. F. (1994).J . h'atl. Cancer Iwt. 86, 695-699. Yokota, J., and Sugimura, 1.(1993). FASEB 7, 920-925. Yokota, J., Wada, M., Shimosato, Y., Terada, M., and Sugimura, T. (1987).Proc. Natl. Acad. Sri. C'SA 84, '3252-9256. You, I.., Wang, D., Galati, A. J., Ross, J. A,, Mass, $1. J., Nelson, G. B., Wilson, K. H., Amin, S., Stoner, J. C . . Nesnow, S., and Stoner, G. D. (1994). Carcinogenesic 15, 2613-2618. You, M., Candrian, U., Maronpot, R. R., Stoner, G. D., and Anderson, M. W. (1989). Proc. Natl. Acad. Sci. USA 86, 3070-3074. You, M., Wang, Y., Stoner, G., You, L., Maronpot, R., Reynolds, S. H., and Anderson, M. (1992a). Proc. Natl. Acad. Scz. U S A 89, 5804-5808. You, M., Wang, Y., Lineen, A. M., Gunning, W. T.. Stoner, G. D., and Anderson, M. W. (1992b). Carrinogenes-tr 13, 1583-1586. You, M., Wang. Y., Nash, B., and Stoner, G. D. (1993). Carcinogenesis 14, 1247-1249. Zhang, Y., Proenca, R., Maffei. M., Barone, M., Leopold, L., and Friedman, J. M. (1994). Nature 372, 425-43 1 .

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS Gianluca Gaidano’ and Riccardo Dalla-Favera2 1Laboratoriodi Medicina e Oncologia Molecolare, Dipartimento di Scienze Biomediche e Oncologia Umana, Universith di Torino, Ospedale San Luigi Gonzaga, Turin, Italy, and 2Division of Oncology, Department of Pathology, College of Physicians & Surgeons of Columbia University, New York, New York 10032

1. Epidemiology of AIDS-Related Lymphomas 11. Clinicopathologic Spectrum of AIDS-Related Lymphomas

111.

IV.

’V.

VI.

VII.

A. General Features B. Systemic AIDS-Related Lymphomas C. AIDS-Related Primary Central Nervous System Lymphomas Natural History of AIDS-Related Lymphomas Host Factors Contributing to AIDS-Related Lymphoma Development A. Disrupted Immunosurveillance B. Chronic Antigen Stimulation C. Cytokine Deregulation Role of Viral Infection in AIDS-Related Lymphomagenesis A. Epstein-Barr Virus B. HIV C. Other Viruses Genetic Lesions Involved in AIDS-Related Lymphomas A. Dominantly Acting Oncogenes B. Tumor Suppressor Loci Conclusions: Distinct Pathogenetic Pathways in the Development of AIDS-Related Lymphomas References

I. Epidemiology of AIDS-Related Lymphomas

(Cancer affects more than 40% of all patients with human immunodeficiency virus (HIV) infection, serving both as an immediate cause of death and as a source of great morbidity (Bernstein and Hamilton, 1993; Levine, 1993; Peters et al., 1991). Non-Hodgkin’s lymphoma (NHL) is the second most frequent cancer associated with AIDS after Kaposi sarcoma (KS) (Bernstein and Hamilton, 1993), and in some AIDS risk groups, namely, the hemophiliacs, NHL overrates KS, representing the most common AIDS-associated neoplasia (Ragni et al., 1993). Though NHL cases in association with HIV infection had been described since 1982 (Ziegler et al., 1982), it was not until 1985 that the Centers for

113 A1)VANCES I N CAKCEX RESEAKC€I, VOL. 67

Copyright 0 1995 hy Academic Press, Inc All rights of reproduction in any fomm reserved.

114

G l A N L G C A CAIDAlSO A N D RICCAHDO DALLA-FAVERA

Disease Controi (CDC) recognized NHL as an AIDS-defining illness (Centers for Disease Control, 1985). T h e incidence of AIDS-related nonHodgkin’s lymphoma (AIDS-NHL) ever since has continued to rise (Bernstein et al., 1989; Hardy et al., 1988). Although at present AIDS surveillance data indicate that NHL occurs in approximately 3% of all AIDS patients, with a relative risk oscillating between 60 and 100 (Beral et al., 1991; Biggar and Robkin, 1992; Casabona et ul., 1991; Italian Cooperation Group for AIDS-related Tumors, 1988; Rabkin et al., 1991; Ross et ul., 1985; Serraino et al., 1992), the real proportion of AIDS-NHL is substantially higher, since immunosurveillance data do not include IVHL diagnosed late in the course of AIDS or at autopsy (Cremer P t al., 1990; Kaplan et at., 1989; Klatt, 1988; Loureiro et 01.. 1988; Monfardini et al., 1990; Wilkes et al., 1988). The introduction of long-term antiretroviral therapy and improvement in supportive therapy, leading to longer life expectancies of AIDS patients, have led to a further increase in AIDS-NHL frequency (Fischl et al., 1987; Gailetol., 1991; Mooreetal., 1991; NCI &C:DC, 1991; Pludaetal., 1990). Various epidemiologic features distinguish AIDS-NHL from AIDSrelated KS (AIDS-KS). First, AIDS-IVHL is a relatively late event in AIDS natural history, whereas AIDS-KS frequently is the presenting symptom (Bernstein and Hamilton, 1993; Ross et al., 1985; Roithman et a(., 199 1). Second, AIDS-NHL displays a relatively itniforin risk across all HI\’-infected risk groups, whereas AIDS-KS is characterized by a strong association with male homosexuality (Beral et a/., 1991 ; Bernstein and Hamilton, 1993; Biggar and Rabkin, 1992; Jaffe et al., 1983). T h e biological basis for the epidemiological differences observed between AIDS-NHL and AIDS-KS is unknown. Finally, the issue of the association between HI\’ infection and Hodgkin’s disease (€ID) remains controversial. Despite initial suggestions (Bernstein and Hamilton, 1993; Hessl et al., 1992; Ioachim, 1992; Reynolds et nl.. 1993; Serrano et al., 1990), the frequency of HD in AIDS patients does not differ from that in the general population (HamiltonDutoit et al., 1991; Miller et d., 1992). However, the characteristics and presentation of AIDS-related HD (AIDS-HD) are strikingly different f’roni those normally reported among noninfected cases (Carbone et al., 1991; Ioachim et ul., 1985; Pelstring bt al., 1991; Prior et al., 1986; Ree et ul., 1991; Safai P / d., 1992; Schoeppel et a/., 1985), indicating the need for continued attention regarding the possible existence of HIVassociated l!.tiiphoproliferative disorders distinct from NHL and resembling HD.

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

115

II. Clinicopathologic Spectrum of AIDS-Related Lymphomas

A. GENERAL FEATURES T h e first NHL to be reported in association with AIDS were represented by Burkitt’s lymphomas (BL) (Ziegler et al., 1982). Subsequent investigations expanded the spectrum of AIDS-NHL to include diffuse large cell lymphomas and immunoblastic lymphomas, as well as anaplastic lymphomas (Ahmed et al., 1987; Beral el al., 1991; Boyle et al., 1990; Carbone et al., 1991; Chadburn et al., 1993; Di Carlo et al., 1986; Doll and List, 1982; Gill et al., 1985; Hamilton-Dutoit et al., 1991; Ioachim et af., 1985, 1991; Kalter et al., 1985; Karp and Broder, 1991; Knowles et al., 1988; Knowles and Chadburn, 1992; Levine et al., 1984; Levine, 1992; Lowenthal et at., 1988; Raphael and Knowles, 1990; So et at., 1986; Ziegler et al., 1984). Two features are universally recognized as distinctive of AIDS-NHL. First, AIDS-NHL are consistently of B-cell origin. Second, AIDS-NHL are high grade or, more rarely, intermediate grade lymphomas according to the Working Formulation (Non-Hodgkin’s Lymphoma Pathologic Classification Project, 1982). In particular, 8090% of AIDS-NHL patients are diagnosed with high grade NHL, which normally would be expected in approximately 10-1576 of NHL series in the general population (Lukes et al., 1978). T h e detailed pathological classification of AIDS-NEIL has been a matter of controversy and is continuously being remodeled (Table I). Part of the difficulty encountered in classifying AIDS-NHL is attributed to AIDS-NHL morphologic polymorphism, which prevents the full application of uniform criteria like the ones standardized by the Working Formulation (Hamilton-Dutoit et al., 1991; Knowles and Chadburn, 1992). For practical purposes, AIDS-NHL are distinguished between systemic AIDS-NHL, accounting for approximately 85% of AIDS-NHL, and primary lymphoma of the central nervous system (CNS), representing the remaining 15% (Table I; Beral et al., 1991; Karp and Broder, 1991; Knowles and Chadburn, 1992; Levine, 1992). Systemic AIDS-NHL are histologically heterogeneous and recognize three main histologic types (Table I): (a) small noncleaved cell lymphomas (SNCCL); (b) diffuse large cell lymphomas (DLCL); and (c) anaplastic large cell lymphomas (ALCL) (Irwin and Kaplan, 1993; Karp and Broder, 1991; Knowles and Chadburn, 1992; Levine, 1992; von Gunten and von Roenn, 1992). Conversely, AIDS-related primary central nervous system lymphomas (AIDS-PCNSL) tend to display more uniform histology, consistent with the DLCL type (Irwin and Kaplan, 1993; Karp

116

GIANLC'CA G A I D A N O A N D RICCARDO DALLA-FAVERA

TABLE I PAI HOLOGICAL CLASSIFICATION OF AIDS-RELATED LYMPHOMAS AIDS-Related Lymphoma B-Cell Non-Hodgkin's Lymphoma (NHL) Systemic AIDS-NHL Small non-cleaved cell lymphoma (SNCCL) Diffuse large cell lymphoma (DLCL) Large noncleaved cell lymphoma (LNCCL) Large cell immunoblastic plasmacytoid lymphoma (LC-IBPL) Anaplastic large cell lymphoma (ALCL) Primary central nervous system lymphoma (PCNSL) Other T-cell Lymphomas Hodgkin's disease (HD) 0

b

Frequencya (%)

85

30b 70h

15

rare rare

Frequencies as derived fi-om Beral ef 01. (1991) Frequencies within s)ctemic AIDS-NHI..

and Broder, 1991; Knowles and Chadhurn, 1992; Levine 1992; von Gunten and von Koenn, 1992). Apparently, all AIDS risk groups are at equal risk for N HL development, without convincing associations between risk groups and distinct AIDS-NHL subtypes (Irwin and Kaplan, 1993).

B.

SYSTEMIC

AIDS-RELATED LYMPHOMAS

As stated earlier, systemic AIDS-related lymphomas include AIDSrelated SNCCL (AIDS-SNCCL),AIDS-related DLCL (AIDS-DLCL), and BIDS-related ALCL (AIDS-ALCL) (Table I). Systemic AIDS-SNCCL accounts for approximately one-third of systeniic AIDS-NHL (Beral et al., 1991; Berard et d., 1989; Carhone et al., 1991; Haniilton-Dutoit et al., 1991; Hui et a / . , 1988: Ioachim et (il., 1991; Knowles and Chadhurn, 1992; Kaphael et al., 1991). When compared to other AIDS-NHL types, AIDS-SNCCL shows a peak of incidence at a younger age and tends to develop as an earliei- inanifestation of HIV infection with higher mean CD4 counts (Beral et (11.. 1991; Boyle et nl., 1990; Roithinan et nl., 1991). .[he strikingly increased frequency of SNCCL among AIDS patients is unique among inmiunodeficiency settings other than AIDS (Beral et nl., 1991). DLCL, is the second niost coninio~iAIDS-related neoplasm occurring in association Lvith AIDS after AIDS-KS (Beral et al., 1991), accounting f o r approximately two-thirds of systemic AIDS-NHL. The definition of

MOLECULAR PATHOGENESIS OF

AIDS-RELATED LYMPHOMAS

1 17

DLCL is inclusive of two subtypes, large noncleaved cell lymphomas (LNCCL) and large cell immunoblastic plasmacytoid lymphomas (LCIBPL), which are now classified as a single category under the term AIDS-DLCL (Harris et al., 1994). The risk of AIDS-DLCL generally increases as immune function decreases, and AIDS-DLCL patients tend to display more severe immunodeficiency than AIDS-SNCCL cases (Boyle et al., 1990; Kalter et al., 1985; Pedersen et al., 1991; Roithman et al., 1991; Yarchoan et al., 1991). The close dependence of AIDS-DLCL upon immunological disruption is strengthened by the frequency of DLCL in immunodeficiencies other than AIDS (Penn, 1978, 1981, 1988, 1990). ALCL is a distinct NHL type that has been recognized in association with AIDS (Carbone et al., 1991, 1993a,b; Chadburn et al., 1993; Raphael et al., 1991). Similar to ALCL in the immunocompetent host, AIDSALCL are diffuse aggressive lymphomas characterized by strong expression of the CD30 antigen (Stein et al., 1985). The striking morphologic and immunophenotypic similarities between AIDS-ALCL and AIDSHD have led some investigators to postulate that these two neoplasms are closely related and differ mainly in the number of tumor cells and the abundance of cellular infiltrates (Agnarsson and Kadin, 1988; Carbone et d., 1993a; Stein et d., 1985). As a whole, systemic AIDS-NHL display a number of peculiarities when compared to NHL of similar histology arising in the immunocompetent host. The predilection of AIDS-NHL for extranodal sites was considered to be a distinctive feature of AIDS-NHL since their first appearance (Ziegler et al., 1982). Depending on different studies, extranodal presentation ranges between 70 and 90%, predominates in the gastrointestinal tract and the bone marrow, and may frequently involve unusual sites, including body cavities, anus and rectum, heart, adrenal, gingiva and oral cavity, salivary glands, muscle, and other soft tissues and the placenta and product of conception (Carbone et al., 1991; Ioachim et al., 1985, 1991; Kaplan et al., 1989; Knowles et al., 1988; Lowenthal et al., 1988; Monfardini et al., 1990; Pollack et al., 1993). Other peculiar features of systemic AIDS-NHL are the late stage at presentation (stage IV), mostly due to bone marrow involvement, and the presence of severe clinical symptoms (Irwin and Kaplan, 1993; Levine et al., 1991). C. AIDS-RELATED PRIMARY CENTRAL NERVOUS SYSTEM LYMPHOMAS With few exceptions (Carbone et al., 1991), AIDS-related PCNSL (AIDS-PCNSL) account for approximately 15-20% of AIDS-NHL (Ta-

118

GIANLUCA G A I D A N O A N D RICCARDO DALLA-FAVERA

ble I ; Baunigartner et a!., 1990; Beral et al., 1991; Bernstein et al., 1989; Di Carlo et al., 1986; Formenti et al., 1989; Ioachim et al., 1985; Loureiro et al., 1988; Lowenthal et al., 1988; Ziegler et al., 1984). Since, in the immunocompetent host, PCNSL comprise 1% of all NHL (O’Neill and Illig, 1989), it is estimated that the relative risk for AIDS-PCNSL is at least 1000 (Beral et al., 1991).T h e usual presentation of AIDS-PCNSL as single tumor masses situated deep in the cerebral white matter prevents an early diagnosis in most cases (Ioachini et al., 1991), and as many as 25% of AIDS-PCNSL are diagnosed only during autopsy (Gill et al., 1985; Lowenthal et a/., 1988; So et al., 1986; Wilkes et al., 1988). AIDS-PCNSL are histologically homogeneous, being primarily represented by DLCL, namely, its LC-IBPL variant (Gill et al., 1985; Goldstein et al., 1991; Knowles and Chadburn, 1992; Levine et nl., 1991; Raphaelet al., 1991; Roithman et nl., 1991). PCNSL of similar histology are also detected in congenital or acquired immunodeficiencies other than AIDS (Frizzera e f al., 1980; Hoover and Frdumeni, 1973; l’enn, 1978, 1981, 1988, 1990). AIDS patients developing AIDS-PCNSL tend to be at a far advanced stage of HIV infection, with a history of AIDS prior to lymphoma in the overwhelming majority of cases and profoundly disrupted immune function and the poorest outlook among AIDS-NHL patients (Levine et al., 1991; Pluda et al., 1990, 1993; Yarchoan et al., 1991).

I l l . Natural History of AIDS-Related Lymphomas The development of B-cell NHL in the context of AIDS often is preceded by symptoms such as polyclonal hypergammaglobulinemia and persistent generalized lymphadenopathy (PGL), which indicate the presence of chronic B-cell stimulation and expansion (Knowles and Chadburn, 1992; Mathur-Wagh et al., 1984; Metroka r’t al., 1983). These observations have been taken to suggest that a pathogenetic relationship may exist between B-cell hyperplasia and the development of B-cell NHL in HIV-infected patients (Fig. 1). Early in the AIDS-NHL era, a correlative imniuiiophenotypic and molecular genetic analysis of the hyperplastic lymphadenopathy (i.e., PGL) associated with HIV infection was perfortned in an attempt to investigate its relationship with AIDS-NHL (Pelicci et al., 1986a). This study revealed that a substantial fraction of morphologically benign, apparently polyclonal lymph nodes collected from HIV seropositive individuals displayed one or more discrete immunoglobulin gene rearI-angenient bands, indicating the presence of one or more B-cell clonal expansions. On the basis of these results, it was proposed that the hyperplastic lymph nodes of- H IV-infected patients often contain clonally ex-

POLY CLONAL

MONOCLONAL

OLIGOCLONAL

I

I

I I DISRUPTED IMMUNOSURVEILLANCE CHRONIC ANTIGEN STIMULATION DISRUPTION OF CYTOKINE NETWORKS EBV INFECTION

I ACCUMULATION OF GENETIC LESIONS

FIG. 1. Clonal progression of AIDS-NHL. In the initial stages, host predisposing conditions, including immunosurveillance alterations, chronic antigen stimulation, deregulation of cytokine networks, and EBV infection, favor the development of a polyclonal to oligoclonal B-cell hyperplasia known as persistent generalized lymphadenopathy (PGL). The polyclonal nature of B cells in the figure is indicated by different cytoplasmic patterns. With time, B-cell oligoclonal expansions arise within the context of the PGL (in the figure, the clone identified by a white cytoplasmic pattern is the one emerging). In the second phase, the accumulation of genetic lesions within a single clone (identified by the white cytoplasmic pattern) leads to the development of a true monoclonal AIDS-NHL.

120

CIANLUCA G A I D A N O A N D RICCARDO DALLA-FAVERA

panded B-cell populations not identifiable by morphologic examination or immunophenotypic analysis. This study also demonstrated that the clonal B-cell expansions found in PGL lymph nodes were devoid of genetic lesions typical of AIDS-NHL (see Section VI), suggesting that they were composed of nontransformed B cells representing putative precursors to AIDS-IVHL (Pelicci et nl., 1986a). In contrast to the oligoclonal representation of PGL lymph nodes, AIDS-NHL are consistently monoclonal (Ballerini et al., 1993; Gaidano et al., 1993, 1994; Pelicci et al., 1986a; Subar Pt al., 1988) and are characterized by the presence of genetic lesions, including c-MYC rearrangements, $53 mutation, BCL-6 truncations, and EBV infection (Ballerini et al., 1993; Gaidano et al., 1993, 1994; Subar et al., 1988). The next sections will outline a pathogenetic model based on the clonal progression of AIDS-associated lymphomagenesis and will summarize the main biological and genetic alterations contributing to the development of PGL and evolution into AIDS-NHL. Despite the general validity of this model, it is important to note that some AIDS-NHL may deve1o.p in the absence of a preceding PGL phase (Knowles and Chadburn, 1992) and that, since genetic lesion accumulation is a stochastic event, only a proportion of PGL cases will progress to AIDS-NHL.

IV. Host Factors Contributing to AIDS-Related Lymphoma Development A. DISRUPTED IMMIUKOSUKVEILLANCE

'The close relationship between decreased immunosurveillance and increased risk of lymphoma has long since been postulated from the case of congenital primary inimunocleficiency syndromes (Gatti and Good, 1971; Kersev et nl., 1973; Penn, 1978; Peterson pt ul., 1964). Iatrogenic inimunosupbressioii further adclecl to this concept and allowed the definition that the deeper the iiiiinuIiosuppressioii, the higher the risk of lyniphonia development (Swinnen P / al., 1990). On these grounds, it came as n o surprise that the AIDS outburst might be paralleled by an epidemic of NHL. What was and remains an intriguing issue, however, is the histological heterogeneity of AIDS-NHL, which include both AIDSDLCI, and .4IDS-Sh'CCL. Only AIDS-DLCL, in fact, develops in congenital and iatrogenic immunodeficiencies. T h e direct contribution of disrupted immunosurveillance per se to AII>S-related lymphomagenesis has been substantiated by animal models, clinicoepidemiological investigations, and in iiitw studies (Table 11). Yriniates inf'ected with the simian immunodeficiency virus (SIV) develop

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

121

TABLE I1 HOSTFACTORS PREDISPOSING TO LYMPHOMA DEVELOPMENT IN THE CONTEXT OF AIDS Host factor

Clinicoexperimental evidence

Disrupted immunosurveillance

Animal models develop NHL after infection with HIV homologues Increasing NHL relative risk with decrease of CD4 counts Low levels of T-TIL infiltrating AIDS-NHL biopsies Selective deficit of cytotoxicity anti-EBV-infected B cells

Chronic antigen stimulation

Polyclonal hypergammaglobulinemia of HIVinfected individuals Some AIDS-SNCCL produce autoantibodies Somatic hypermutation of immunoglobulin genes hypervariable regions utilized by AIDS-SNCCL

Cytokine deregulation

Spontaneous release of IL-6 by EBV-infected circulating B cells Marked expression of IL-6 receptors by AIDSNHL Predictive value of IL-6 serum levels with respect to NHL development High levels of IL-10 production by AIDS-SNCCL

malignant lymphomas in a high proportion of cases 5-15 months after viral inoculation (Feichtinger et al., 1990). All lymphomas observed in this animal model display a B-cell phenotype and high grade morphology. Consistent with the absence of HIV particles from human AIDSNHL, SIV-associated lymphomas are devoid of SIV genomes (Feichtinger et al., 1990). 'The most convincing evidence for the role of immunosuppression in the pathogenesis of AIDS-NHL comes from clinicoepidemiological investigations, which have determined the risk of AIDS-NHL development according to peripheral blood CD4 levels. Several investigators have determined that the greatest risk for AIDS-NHL development occurs when CD4-positive lymphocyte counts are <50/p1 (Levine et al., 1991; Moore et al., 1991; NCI 8c CDC, 1991; Pluda et al., 1990, 1993; Yarchoan et al., 1991). The association between low CD4 counts and AIDS-NHL hazard is particularly striking in the cases of AIDS-DLCL and AIDS-PCNSL, whereas AIDS-SNCCL also may be observed in the presence of a substantially preserved immunity (Kalter et al., 1985; Levine et al., 1991; Pedersen et al., 1991; Pluda et al., 1993; Roithman et al.,

I22

GIANLUCA C A I D A N O A N D KICCARDO DAL1.A-FAVERA

1991). 'Time, i.e., prolonged exposure to immunosuppression, appears to be another critical factor. Thus, for example, in one study the probability of anv patient (independent of hidher CD4 counts) developing AIDS-NHL after 24 months of antiretroviral therapy with AZT was 12%, which increased to 29.2% after 36 months (Pluda et al., 1993). Taking into consideration both CD4 counts and exposure time to immunosuppression, Pluda et al. estimated that the risk of a patient developing AIDS-NHL after having less than 50 CD4 cells/pI for 24 months is 26.9%,which is more than double the value obtained when AIDS patients are not selected for their CD4 counts (Pluda rt d.,1993). Since dideoxynucleosides are animal carcinogens (Ayers, 1988), an obvious question is whether the increased AIDS-NHL risk was caused by the mutagenic effect of the drug. However, several lines of evidence ruled out this possibility on the basis of animal and clinical studies (de Miranda et al.. 1990; Pluda et al., 1993; Volderbing rt al., 1990). Besides the role of generalized immune deficiency, XIDS-NHL pathogenesis is also affected by local 111 situ failure of T-cell response. The role o f tumor infiltrating T lymphocytes (T-TIL) in tumor containment is well established, and the magnitude of 'r-TIL response in B-cell NHL biopsy specimens has been suggested as an independent predictor of clinical outcome (Ramsay et d.,1988). When AIDS-DLCL are compared to DLCL of the immunocompetent host, the former group displays a markedly lower representation of T-TIL in the biopsy specimen and a significantly greater incidence of histological forms with low representation of T-TIL (List et a[., 1993). This is in line with the profound functional defects of CD8-positive T cells in persons infected with HIV (Fauci et al., 1991). Finally the contribution of immunosuppression per se to AIDS-NHL pathogenesis is also documented by the fact that one of the AIDSassociated immunological defects selectively impairs the imrnunosurveillance against EBV-infected B cells, which are present in increased numbers in the patients' blood and lymphoid organs and may be responsible for minor clonal B-cell expansions that precede neoplastic transformation (Birx et al., 1986). Although never formally demonstrated, it seems logical to believe that these clorial expansions are in fact the precursors of AIDS-NHL. B.

(:HHONIC

ANTIGEN STIMrLATION

HIV infection is characterized by a gradual though progressive deregulation of the immune system and by significant, chronic antigenic

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

123

stimulation of B cells (Martinez-Maza et al., 1987; Yarchoan et al., 1986), leading to polyclonal hypergammaglobulinemia and florid B-cell hyperplasia within enlarged, reactive PGL lymph nodes (Fauci et al., 1991; Knowles and Chadburn, 1992; Lane et al., 1983; Nath et al., 1987). HIV per se is one direct cause of polyclonal B-cell activation, which may be operative along with HIV-induced cytokine release (Schnittman et al., 1986). Indeed, the frequent oligoclonal bands that accompany AIDS-related hypergammaglobulinemia display anti-HIV reactivity (Ng et al., 1988, 1989; Papadopoulos et al., 1988). Other antigens contribute to polyclonal B-cell activation, including both environmental and selfantigens (Fauci et al., 1991). The role of self-antigens might be particularly relevant in light of the frequent autoimmune manifestations associated with AIDS (Kopelman and Zolla-Pazner, 1988). T h e crucial role of antigens in normal B-cell development is a welldocumented notion of basic immunology (Berek and Milstein, 1987; Ikematsu et al., 1993; Ueki et al., 1990). In NHL of the immunocompetent host, antigen stimulation may play a role in B-cell expansion and selection preceding and/or associated with lymphoma development (Hahler and Levy, 1992; Friedman et al., 1991; Levy et al., 1988; Zelenetz et al., 1992). Recently, the contribution of chronic antigen stimulation to AIDS-NHL development was directly tested by analyzing the antigenic reactivity and genetic structure of the immunoglobulin gene variable regions of heavy (V,) and light chains (V,) utilized by AIDS-NHL (Riboldi et al., 1994). The assessment of a potential role for antigen in B-cell lymphomagenesis first of all entails the definition of the specificity of the tumor-derived antibody. Homogeneous in vitro growing AIDSSNCCL cell populations (Gaidano et al., 1993) were found to produce high-affinity IgM monoclonal antibodies (mAb), which were self-reactive with the i determinants on red blood cells and behaved as cold agglutinins (Riboldi et al., 1994). The IgM mAb V, segment was represented by the V,4.21 gene, which is also frequently utilized by autoreactive B-cell clones in other diseases (Riboldi et al., 1994; Silberstein et al., 199 1). It is well-known that one of the hallmarks of antigenic stimulation and selection of a B-cell clone is the accumulation of somatic mutations in the immunoglobulin complementarity determining region (CDR), which leads to increased affinity of the antibody for its antigen (Berek and Milstein, 1987). Sequence analysis of the V, and V, genes utilized by AIDS-SNCCL producing anti-i mAb showed that the CDR were the site of selective clustering of somatic mutations, the nature of which was consistent with mutations selected by antigen stimulation (Riboldi et al.,

124

GIAKLUCA GAI DANO AND RICCARDO

DALLA-FAVERA

1994). Altogether, these data provide indirect evidence that chronic antigen stimulation is involved in AIDS-SNCCL pathogenesis (Table 11). Studies addressing its role in other AIDS-NHL types are in progress. An antigen-driven process of clonal selection may play a role in the emergence and/or expansion of a neoplastic B-cell clone. It is unclear precisely at which stage chronic antigen stimulation exerts its influence in AIDS-SNCCL development. T h e putative self-antigen stimulation driving B-cell clonal expansion and selection may precede the occurrence of genetic lesions or, conversely, may follow it. In the first case, that is, chronic antigen stimulation preceding the occurrence of genetic lesions, chronic antigen stimulation would drive a genetically normal B cell to hyperproliferation, thus increasing the probability of genetic accidents. In the second scenario, that is, chronic antigen stirnulationfollowm g the occurrence of genetic lesions, antigen might act as a rnitogen contributing to the expansion of a genetically abnormal neoplastic cell, which necessitates an external growth stimulus.

C. CYTOKINE DEREGULATION Disruption of normal cytokine networks is a key feature of HIV infection (Fauci et al., 1991). Yet, the contribution of cytokines to AIDSrelated lymphoniagenesis is a relatively unexplored field. Theoretically, the Contribution o f cytokines may be twofold, both indirect and direct (Table 11). On one hand, cytokine deregulation may help in maintaining or severing HIV infection, which in turn would further worsen immune function and thus facilitate AIDS-IVHL development. On the other hand, numerous cytokines induced by HIV regulate B-cell proliferation, and their deregulation through paracrine or autocrine loops may induce or sustain the growth of B-cell malignancies. 1.11~ ~ i t r o€3, cells from HIV-infected individuals with hypergammaglobulineniia constitutively secrete high levels of TNF-a and interleukin (1L)-6 in the absence of exogenous stimuli, whereas B cells from seronegative persons require in uitro stimulation (Fauci et al., 1991; Rieckmann et al., 1991). In AIDS patients, the intense antigenic exposure o f H cells in z h o directly activates IL-6 and T N F - a production. In turn, IL-6 and TNF-a induce HIV expression, thus maintaining viral infection (Fauci rt al., 1991). .4mong the many cytokines regulating B-cell growth, IL-6 and IL- 10 have been documented to be involved in AIDS-NHL growth. IL-6 plays ;I role in the development of B-cell tumors of the immunocompetent host (Kawano et al., 1988; Kishimoto, 1989; Levy et nl., 1991; Schwab et d.,1991; Tohyama P t al., 1990) and potentiates the tumorigenicity of

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

125

E:BV-infected B cells (Scala et al., 1990; Tanner and Tosato, 1991). In the AIDS context, HIV directly stimulates IL-6 production from monocytes and macrophages. IL-6 in turn promotes the chronic proliferation of activated (e.g., by EBV) B cells, thereby driving immunoglobulin synthesis and causing the nonspecific hypergammaglobulinemia commonly seen in early HIV infection (Amadori et al., 1991; Birx et al., 1990; Elmilie et al., 1990; Yarchoan et al., 1986). In addition to HIV-infected monocytes, IL-6 is also produced by activated B cells. The combination of these mechanisms causes the relatively high levels of IL-6 generally observed in HIV patients (Breen et al., 1990). The role if IL-6 as a host factor predisposing to AIDS-NHL development is corroborated by the observation that IL-6 serum levels in AIDS patients may be predictive of subsequent NHL development (Pluda et al., 1993). Once the lymphoma is fully established, its continuous growth and expansion may be driven by IL-6 through paracrine loops (Emilie et al., 1992a). This seems particularly true in AIDS-DLCL containing immunoblasts, whereas IL-6 expression is consistently absent from AIDS-SNCCL biopsies (Emilie et al., II 992a). In AIDS-NHL, endothelial cells and macrophages intermingled with the tumor would release IL-6, which in turn would act on IL-6 receptors expressed at high levels by AIDS-NHL (Emilie et al., 1992a). In contrast to the paracrine role of IL-6, IL-I0 is thought to act through an autocrine loop. IL-10 is a pleiotropic cytokine with striking homology to BCRF1, an EBV protein (Vieira et al., 1991). IL = 10, but not BCRF1, is a potent B-cell stimulator (Vieira et al., 1991). EBV-positive AIDS-SNCCL produce high quantities of IL-10 (Benjamin et al., 1992; Emilie et al., 1992b). IL-10 production is a peculiar feature of SNCCL arising in the AIDS context, since IL-10 production is absent from both sporadic and endemic BL cases (Benjamin et al., 1992). Intriguingly, although BCRFl expression is never detected in AIDS-NHL, IL-10 production by AIDS-NHL clusters with EBV-positive cases (Benjamin et al., 1992; Emilie et al., 1992b). T h e role of IL-10 in AIDS-related lymphoproliferations is probably a complex one. The B-cell-differentiating act ivity of IL-10 could be responsible for enhanced immunoglobulin secreI ion, and, consequently, the hypergammaglobulinemia of AIDS patients (Rousset et al., 1992). On the other hand, IL-10 is also a viability factor for B cells (Go et al., 1990). By inhibiting apoptosis, IL-10 may facilitate .4IDS-NHL expansion. Finally, by inhibiting the T-cell production of [L-2 and yIFN (de Waal Malefyt et al., 1991; Fiorentino et al., 1991), IL-I0 may impair the immunosurveillance against EBV-infected B cells, .which are present at high levels in AIDS patients. The pattern of cytokine expression of systemic AIDS-NHL shows some peculiarities when compared to the pattern observed in NHL of

126

GIANLUCA G A I D A N O AND RICCAKDO DALLA-FAVERA

similar histology arising in the imniunocompetent host. Preliminary data from our laboratory suggest that the expression of IL-6, IL-7, and IL-13 is peculiar to SNCCL arising in the AIDS context, whereas it is generally absent from SNCCL of the immunocompetent host (our unpublished observation). It is possible that the distinct pattern of cytokine expression observed in AIDS-NHL is due to the activation of gene expression pathways proper of AIDS-NHL and putatively corresponds to specific genetic lesions. V. Role of Viral Infection in AIDS-Related Lymphomagenesis A. EPSTEIN-BARK VIRUS Ever since the outbreak of the AIDS-NHL epidemic, the role of Epstein-Barr virus (EBV) has been subject to a great deal of study and speculation. First, serological data and analysis of viral shedding in saliva have demonstrated that more than 90% of HIV-infected individuals were also EBV infected and that reactivation of EBV infection was a common occurrence in AIDS (Katz et al., 1992; Peiper et al., 1990; Sumaya et al., 1986). Second, the mean number of circulating EBV-infected B cells capable of spontaneous growth in uitro is increased in AIDS patients (Birx P t nl., 1986). Finally, the association between the presence of EBV and the development of B-cell lymphoproliferative disorders is well recognized (Epstein et al., 1964; Gaidano and Dalla-Favera, 1993; Kieff and Liebowitz, 1989; Klein, 1992; Purtilo ef al., 1992; zur Hausen et al., 1970). The frequency of EBV infection, however, has represented a matter of controversy due to differences in the technical approaches utilized for viral detection. T h e frequent presence of polyclonal B-cell expansions infected by EBV and contaminating AIDS-NHL tissues mandates the use of experimental approaches allowing selective detection of EBV genomes/transcripts in tumor cells (Birx el al., 1986). At present, there is uniform agreement that the frequency of EBV infection varies dramatically according to AIDS-KHL histology (Ballerini et al., 1993; Bashir et nl., 1989; Carbone eta/., 1993a; Chadburn et al., 1993; De Angelis et al., 1992; Gaidano and Dalla-Favera, 1992a; Gaidano et al., 1993; Guarner et ul., 1991; Hamilton-Dutoit et al., 1991, 1993; MacMahon et al., 1991; Meeker et al.. 1991; hlorgello, 1992; Keri et at., 1991; Paulus et al., 1993; Pedersen et al., 1991; Ragni et al., 1993; Shibata et al., 1993; Subar et al., 1988).Thus, the rate of EBV infection ranges from approximately 30% in systemic AIDS-SNCXL to 100% in AIDS-PCNSL (Table 111). Systemic

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

127

TABLE I11 GENETIC LESIONS AND VIRAL INFECTIONIN AIDS-NHL" Tumor supressor genes

Dominantly acting oncogenes

Viruses

ECL-1

ECL-2

RAS

p53

RE1

EBV

HIV

(%I

(%I

(%I

(%I

(%I

(%I

(%I

(%I

(XI

AIDS-SNCCL

100

-

-

-

15

60

-

30

-

IDS-DLCL AIDS-LNCCL AIDS-LC-IBPL AIDS-ALCL

20 20 NA

20 20 -

-

NA

30

-

AIDS-PCNSL

NA

NA

-

NA

100

C-MYC ECL-6

-

15 1 5 NA -

-

NA

-

-

9 0 90

-

* When positive, the frequency is given. -, gentic lesion not involved; NA, not available.

AIDS-DLCL shows an infection rate oscillating between 70 and 80% (Table 111). Interestingly, among AIDS-DLCL variants, a higher frequency of EBV infection is detected in AIDS-LC-IBPL (90%) than in AIDS-LNCCL (30-50%) (Table 111). The few cases of AIDS-ALCL and AIDS-HD characterized have shown an extremely high frequency of EBV infection (Carbone et d., 1993a,b; Chadburn et al., 1993; Hamilton-Dutoit et d.,1993). As it is the most frequent genetic lesion detected in AIDS-NHL, several investigators have tried to define the clinical relevance of EBV infection in AIDS-NHL. Among systemic AIDS-NHL, several investigations have attempted to define whether EBV infection confers peculiar features to the lymphomas. Thus, Hamilton-Dutoit et al. reported no morphological differences between EBV-positive and EBV-negative NHL within each histologic group, whereas they noticed that EBV infection of systemic AIDS-NHL was particularly associated with extranodal disease, with the frequency of viral infection of extranodal AIDS-DLCL approaching that of AIDS-PCNSL (Hamilton-Dutoit et al., 1993). In addition, Pedersen et al. showed that EBV-infected, systemic AIDS-NHL had significantly lower CD4 cell counts and lower CD8 counts than patients with EBV-negative AIDS-NHL (Pedersen et ad., 1991). With respect to AIDS-PCNSL, the diagnosis of which is a critically demanding issue for clinicians, the consistent association with EBV infection (Hamilton-Dutoit et al., 1993; MacMahon et al., 1991) has been regarded as a valuable tumor marker for clinical usage, since shedding of EBV DNA into the cere-

I28

C;I..ZNLL'CA G A I D A N O A N D KICCAKDO DALLA-FAVERA

brospinal fluid is detectable before the lymphoma lesions are seen by C T scan or M R I (Cinque ~t al., 1993). 'Ihe demonstration of any virus within a tumor does not amount to proof that it is directly involved in its pathogenesis. In the case of EBV, however, various observations suggest its pathogenetic role. First, EBV is capable of transforming B cells 272 uiiro (Kieff and Liebowitz, 1989). Experiments in the SCID mouse model have corroborated the transf'orniing activity of EBV in zjiuo also, by demonstrating that the injection of EBV-infected B cells into animal recipients causes B-cell tumors within weeks (Rowe rt ctl., 1991). These human SCID tumors histologically resemble the EBV-positive DLCL that develop in immunosuppressed patients and are frequently devoid of genetic lesions other than EBV. Another feature supporting the pathogenetic role of EBV infection in AIDS-NHL is the consistent monoclonality of infection (Ballerini et al., 1993; Gaidano et al., 1993, 1994; Neri et al., 1991; Shibata et al., 1993). I n the infectious virion, the DiXA molecule is linear and flanked l>y variable nutnbers of terminal repeats at either end of the genome. Upon infection of a B lymphocyte, the formation of circular episonies is mediated by the terminal repeats. Because of the termini heterogeneity, the number o f repetitive sequences enclosed in newly formed episomes represents a constant clonal marker of the infected cell (Brown et al., 1988; Raab-Traub and Flynn, 1986). A single form of fused EBV termini will be detectable in the clonally expanded progeny of a single infected cell, whereas heterogeneous EBV termini will be detected in the case of infectiori of either an already expanded clonal population or a polyclonally infected population. T h e single form of fused EBV termini detected in AIDS-NHL indicates that the lymphoma represents the clonally expanded progeny of a single infected cell (Ballerini et al., 1993; Gaidano et al., 1993, 1994; Neri et al., 1991). This concept has been (:orroborated by the analysis of EBV termini in one AIDS-NHL case, from !%.hid biopsies from multiple sites were available (Shibata et al., 1993). In every site involved, the identical EBV termini rearrangement was detected, suggesting that dissemination of lymphoma occurred primarily through the clonal expansion of a single EBV-infected B cell. Finally, the hypothesis that EBV infection precedes clonal expansion is also supported by the morphologic observation that essentially all tumor cells within ElBV-infected AIDS-KHL biopsies carry the viral genome and express ~ i r a genes l (Hamilton-Dutoit et al., 1993). Further evidence for the pathogenetic role of EBV comes from work by Shibata and colleagues, who documented that EBV is detected in approximatel) 40% of reactive lymph nodes from HIV-infected patients with PGL and that the presence of EBV in the context of PGL is a

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

129

significant risk factor for the future development of lymphoma, with persistence of the EBV genome in the tumor clones (Shibata et al., 1991). Since, however, most HIV-infected individuals with EBV-positive PGL do not develop lymphoma, EBV per se is not sufficient for systemic AIDS-NHL development, and additional genetic events are needed. T h e pattern of expression of the EBV transforming proteins EBNA-2 and LMP-1 indicates distinct pathogenetic pathways in distinct AIDSNHL histotypes. It is well established that EBNA-2 and LMP-1 play central roles in the initiation and maintenance of EBV-induced B-cell growth and proliferation (Kieff and Liebowitz, 1989). EBV strains deleted for EBNA-2 are unable to transform lymphocytes, whereas gene transfer of both EBNA-2 and LMP-1 has oncogenic activity in rodent cell lines and causes marked phenotypic changes in B cells (e.g., upregulation of the activation antigens CD23, LFA-1, LFA-3, and ICAM-1) and overexpression of the BCL-2 gene product (Cohen et al., 1989; Gregory et al., 1988; Henderson et al., 1991; Wang et al., 1985, 1990). Three distinct patterns of EBV latency have been described and termed Latency I (absence of both EBNA-2 and LMP-I), Latency I1 (expression of LMP-1 in the absence of EBNA-2), and Latency 111 (expression of both EBNA-2 and LMP-1) (Kieff and Liebowitz, 1989; Rowe et al., 1992). Latency I and Latency I11 associate with BL and lymphoblastoid cell lines, respectively, whereas Latency I1 is shared by nasopharyngeal carcinoma and HD (Klein, 1992). The differential expression of EBV latency genes in lymphoid cells has been a puzzling issue for a long time; a putative explanation has been suggested by the work of Contreras-Brodin et al., who demonstrated that the expression of the EBV latency genes is partially dependent on the differentiation stage of the host cell (Contreras-Brodin et al., 1991). All three latency patterns are found in AIDS-NHL, though with histological restrictions. Among systemic AIDS-NHL, AIDS-SNCCL is associated in virtually all cases with a Latency I phenotype, whereas AIDSDLCL and AIDS-ALCL tend to display a Latency 111 or a Latency I1 pattern in most cases, with a few cases showing the Latency I pattern (Carbone et al., 1993b; Hamilton-Dutoit et al., 1993). Where expressed, EBNA-2 and LMP-1 appear to be associated with extranodal lymphoma, with expression of cell surface activation antigens (CD23, CD30, CD39, CDw70) (Carbone et al., 1993b; Hamilton-Dutoit et al., 1993). Finally, AIDS-PCNSL are consistently characterized by the expression of both EBNA-2 and LMP-1, mimicking the Latency 111 pattern. Overall, in patients with immune defects, expression of LMP-1 and EBNA-2 is thought to be an important factor in EBV-associated lymphomagenesis (Klein, 1975, 1989a; Pallesen et al., 1991; Young et al., 1989). Because

130

(;I;ZNLUCA GAIDANO A N D RICCARDO DALLA-FAVERA

both LMP- 1 and EBIVA-2 can serve as targets for cytotoxic T cells, their expression induces T-cell immunosurveillance and thus facilitates the control of the neoplastic process in patients with preserved immune function (Moss ei al., 1988; Murray et al., 1988). Conversely, in the setting of profound immunodeficiency, such as during late phases of HIV infection, cytotoxic T-cell function against EBV is strongly impaired (Birx et al., 1986),and the host immune system may become permissive of EBNA-2 and LMP- 1 expression, leading to uncontrolled proliferation of EBV-infected cells. T h e balance between EBNA-2 and LMP-1 imniunogenicity, on one side, and oncogenicity, on the other side, thus would be regulated by the host immunological conditions. It is noticeable that the AIDS-NHL histotypes that are associated with a more disrupted immunosurveillance (i.e., AIDS-DLCL and AIDS-PCNSL) are the ones that tend to express EBNA-2 and LMP-1, whereas AIDSSNCCL, which arises in the presence of a better preserved immune function, consistently down-regulates EBNA-2 and LMP- 1 (Carbone et al., 1993b; Hamilton-Dutoit et al., 1993). The correlation between host residual immunosurveillance and EBV status in AIDS-NHL is further strengthened by restriction of the EBV lytic cycle (in terms of the expression of lytic cycle antigens such as VCA) to AIDS-PCNSL (Bashir et al., 1993). Overall, the role of EBV in AIDS-related lymphomagenesis appears to be tightly regulated by the levels of residual immunosurveillance of the host, as inferred from (a) the relationship between AIDS-NHL infection by EBV and peripheral blood CD4 counts and (b) variability of the EBV expression pattern. In fact, the highest frequency of EBV infection among ‘4IDS-XHL is found in AIDS-PCNSL, which are associated with the lowest degree of host immune function (i.e., lowest CD4 cells levels), followed by systemic AIDS-DLCL, which are associated with CD4 counts significantly i o t w than those of systeniic AIDS-SNCCI, (Pedersen et al., 1991). In addition, the expression of the highly immunogenic EBV transforming antigens EBNA-2 and LMP-1 is restricted to AIDS-NHL cases arising in the context of profoundly disrupted immunity (Carbone p t al., 1993b; Hatnilton-Dutoit et al., 1993). These data indicate that the role of EBV infection is strictly dependent upon the level of immunity against EBV and requires highly permissive immunological conditions.

B. HIV HIV per se has also been regarded as a potential direct etiologic agent of AIDS-NHL. This hypothesis was based on the indirect evidence that EBV-immortalized B-cell lines may be transformed by HIV in vitro (Lau-

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

131

rence and Astrin, 1991; Pauza et al., 1988) and that, when inoculated into SCID o r nude mice, HIV-transformed B cells give rise to lymphoproliferations resembling lymphomas (Laurence and Austrin, 1991). In spite of these experimental studies, most investigations reported that HIV genomic sequences are not present in AIDS-NHL biopsies (Gaidano et al., 1993; Groopman et al., 1986; Pelicci et al., 1986a; Rechavi et al., 1987; Subar et al., 1988). Furthermore, quantitative PCR analysis of AIDS-NHL tissue has shown levels of HIV that would be predicted from the presence of infiltrating T cells, as opposed to actual HIV infection of the B-cell tumor clone itself (Shibata et al., 1989). Previous to this large number of studies ruling out a direct role of HIV in AIDS-NHL development, Herndier et al. reported the integration of HIV-1 in the genome of one case of AIDS-related T-cell lymphoma, composed of CD4positive lymphoblasts expressing p24 antigens (Herndier et al., 1992). Three more cases with similar features have been added to the literature by the same research group (Shiramizu et al., 1994), who also noticed that HIV integration in these cases occurred clonally within the FUR gene 1 kb apart from the c-FESIFPS protooncogene. It is of note that the HIV-positive lymphoproliferations described by Herndier et al. and Shiramizu et al. displayed peculiar morphological and immunophenotypic features in no way consistent with a B-cell origin, which prevented the diagnosis of one of the standard AIDS-NHL histotypes. At present, any claim for a direct role of HIV in AIDS-related lymphomagenesis should be considered with extreme caution. It remains conceivable that HIV plays an indirect role in AIDS-NHL pathogenesis, for example, by inducing cytokine deregulation of the microenvironment (Fauci et al., 1991) or by chronic antigen stimulation by HIV antigens (Amariglio et a/., 1994; N g et al., 1994).

C. OTHERVIRUSES T h e role of viruses other than EBV in AIDS-related lymphomagenesis has also been tested (Table 111). As already pointed out for EBV, the wide cellular heterogeneity of AIDS-NHL biopsies (Knowles and Chadburn, 1992), often containing a substantial proportion of reactive cells in addition to the tumor clone, demands the use of technical approaches capable of determining whether viral sequences or proteins are present within the tumor cells. All studies claiming viral positivities in AIDSNHL but not attaining this technological requirement should be interpreted with caution (Borisch et al., 1991; Torelli etal., 1991). There is no evidence of AIDS-NHL infection by HTLV-I, HTLV-11, HHV-6, or CMV (Chadburn et al., 1993; Gaidano and Dalla-Favera, 1992a; Gaidano

132

GIANLUCA GAIDANO A N D RICCARDO DALLA-FAVERA

et al., 1993; Karp and Broder, 1991; Paulus et al., 1993; Pelicci et al., 1986a; Subar et al., 1988). VI. Genetic Lesions Involved in AIDS-Related Lymphomas

The observation that EBV per se is not sufficient to promote a full in z~ivotransformation of primary human B cells, in association with the lack of' EBV sequences in a large fraction of AIDS-NHL cases, has prompted a detailed search for additional genetic events contributing to the development of AIDS-related lymphomas. It is now clear that, like most human cancers, AIDS-IVHL display a number of genetic lesions involving dominantly acting oncogenes as well as tumor suppressor genes (Gaidano and Dalla-Favera, 1992b). A. DOMINANTLY ACTINGONCOGENES

Several dominantly acting oncogenes are known to be involved in AIDS-related lymphomagenesis through chromosomal translocation and/or point mutation. These include the c-MYC transcription factor, BCL-6 (a novel zinc finger protein), and the RAS family genes. 1. c-hlYC

Since the initial phases of the AIDS-NHL epidemic. cytogenetic studies had revealed substantial similarity between AIDS-SNCCL and BL of the immunocompetent host based on the presence of chromosomal translocations affecting band 8q24, the site of the c-MYC protooncogene (Chaganti et al., 1983; Klein, 1989b; Klein and Klein, 1985; Magrath et al., 1983; Wang-Peng et al., 1984). These data were confirmed by subsequent c) togenetic investigations (Bernheim and Berger, 1988; Gaidano et al., 1993; Ganser et al., 1988; Roncella et al., 1993) and expanded by molecular analysis of the C-~CIYClocus in AIDS-NHL (Ballerini et al., 1992, 1993; Bhatia et al., 1994; Delecluse et al., 1993; Gaidano ~t al., 1993, 1994a; Groopman et al., 1986; Haluska et al., 1989; Meeker et al., 1991; Pelicci et al., 1986a; Saglio et al., 1993; Subar et al., 1988). Activation of c-h.lYC by chromosomal translocation and/or point mutations are associated with 100% of AIDS-SNCCL, whereas it is restricted to a minority (approximately 20%) of systemic AIDS-DLCL (-Pdble 111; Ballerini et al., 1992, 1993; Bhatia et al., 1994; Delecluse et al., 1993; Gaidano et al., 1994; Subar et al., 1988). Data regarding the role of c-MYC in AIDS-YCKSL and systemic AIDS-ALCL either are not available or are too preliminary t o draw any firm conclusion (Chadburn

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

133

et al., 1993). With respect to AIDS-DLCL carrying c-MYC rearrangements, Delecluse et al. have proposed that they may in fact represent a subset of AIDS-SNCCL that have adopted a large cell immunoblastic morphology in the context of disrupted immunosurveillance (Delecluse et al., 1993). This hypothesis is interesting for several reasons. First, the existence of AIDS-related lymphomas with morphological features intermediate between those of AIDS-SNCCL and AIDS-DLCL has been confirmed by several pathologic reports (Carbone et al., 1993b, 1994; Raphael et al., 1991). Second, “conventional” SNCCL cells, particularly if infected by EBV, frequently undergo an immunoblast-like morphological transition during serial passages in in vitro culture. This morphological transition is accompanied by immunophenotypic variations, as well as by a change in the expression pattern of EBV latent antigens (Rooney et al., 1986; Rowe et al., 1986a,b, 1987). Finally, c-MYC-positive AIDSDRCL tend to display clinical features typical of both AIDS-SNCCL and A IDS-DLCL. With AIDS-SNCCL they share a preferential association with a preexistent PGL, while the host immunodepression is usually extremely advanced, as typically seen in AIDS-DLCL (Delecluse et al., 1993). As a unifying hypothesis for the histogenesis of c-MYC-positive A [DS-DLCL, Delecluse et al. suggested that severe perturbation of the immune system would act as a permissive factor for the histological switch of AIDS-SNCCL to large cell morphology while maintaining the genetic hallmark of AIDS-SNCCL, that is, c-MYC activation (Delecluse et al., 1993). As an alternative hypothesis, we have suggested that AIDSDLCL is characterized by a certain degree of pathogenetic heterogeneity anid that the c-MYC activation observed in 20% of the cases may in fact represent one of multiple, mutually exclusive, genetic pathways active in A [DS-DLCL pathogenesis (Gaidano et al., 1994). It is well established that different mechanisms lead to c-MYC deregulation in lymphoid neoplasia (Dalla-Favera, 1993; Gaidano and DallaFavera, 1995). These include (a) gross truncations within or around the c-MYC locus, (b) point mutations or small abnormalities in the first intron-first exon regulatory regions, and (c) amino acid substitutions in the second exon coding region. Three types of reciprocal chromosomal translocations have been shown to involve the c-MYC locus on chromosome 8q24 and one of the immunoglobulin (Ig) loci, namely, Ig,, Ig,, artd Ig,. In 80% of the cases, t(8;14)(q24;q32) is detectable in which break points located centromeric to c-MYC lead to its translocation into the Ig, locus on chromosome 14q32. In the less frequent variant t(2;8)(pll;q24) (15%) and t(8;22)(q24;qll)(5%) translocations, an IgL locus is translocated to the c-MYC locus, which remains on chromosome 8. While fairly homogeneous at the microscopic level, these chromosom-

134

GIANLCTCA G A I D A N O A N D RICCARDO DALLA-FAVERA

a1 recombinat ions are characterized by remarkable heterogeneity when dissected molecularly. T h e first distinguishing feature is the position of the break points, which are located 5' and centromeric to c-MYC in the t(8;14) translocation, while they are located 3' to c-MYC in the variant t(2;8) and t(8;22) translocations (Dalla-Favera et al., 1982, 1983; Davis et ul., 1984; Hollis et al., 1984; Taub et al., 1982).Further molecular heterogeneity can be found among t(8; 14) translocations, which can be divided into two main types depending on the position of the chromosomal break points on chromosomes 8 and 14 (Haluska et al., 1986; Neri et al., 1988a; Pelicci et al., 1986b; Shiramizu et al., 1991). One type, found more frequently in endemic type BL, involves sequences on chromosome 8 at an undefined distance (> 100 kb) 5' to the c-MYC locus and sequences on chromosome 14 within o r in proximity to the Ig, joining (J) regions. T h e second type, found in most NHL of the immunocompetent host including sporadic BL, involves sequences within or immediately 5' (<3 kb) to the c-MYC gene on chromosome 8 and sequences on chromosome 14 within the IgH switch (S) region. T h e molecular mechanisms leading to c-MYC activation in AIDS-SNCCL are similar to sporadic BL, as opposed to endemic BL, in many aspects, including the analogous break point location on both chromosomes 8 and 14, as well as a comparable frequency of association between c-MYC activation and EBV infection (Ballerini et d., 1993; Subar rt al., 1988). T h e similarity between AIDSSNCCL and sporadic BL is further corroborated by the immunophenotypic observation that both AIDS-SNCCL and sporadic BL express CDlO and lack CD21, whereas endemic BL is usually CDlO negative and CD21 positive (Gaidano et al., 1993; Magrath, 1990). c-MYC expression is tightly regulated during normal cell proliferation and differentiation by transcriptional and posttranscriptional mechanisms. Experimental observations have indicated that this regulation is lost after chromosomal translocation and that translocated c-MYC alleles may be expressed constitutively under the control of Ig regulatory elements (Dalla-Favera et al., 1993; Eick and Bornkamm, 1989; Grignani et al., 1990; Haluska et al., 1987). However, the precise mechanism by which the expression of translocated c-MYC genes is deregulated and the precise role of Ig regulatory domains remain unclear. One additional feature of translocated c-MYC genes that may be relevant to explain the mechanism involved in its deregulation is represented by the consistent presence of structural alterations in the 5' regulatory domain, either through chromosomal decapitation of the c-MYC first exon or through small mutations in a 400-bp region spanning the first exon-first intron junction (Cesarman et al., 1987; Pelicci et al., 1986b). Finally, a

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

135

novel mechanism of c-MYC activation is represented by the recent identification of amino acid substitutions of the c-MYC coding sequence within its second exon (Bhatia et al., 1993).This newly identified mechanism of c-MYC activation appears to be operative among various NHL types and also in AIDS-NHL (Bhatia et al., 1994). The functional role of these mutations consist of redirecting the c-MYC properties of heterodimerization with its partner proteins (see the following; Gu et al., 1994). Substantial experimental evidence documents that the constitutive expression of a normal c-MYC protein can influence the growth and differentiation of B cells in vitro and in vivo. In vitro, the expression of c-MYC oncogenes transfected into EBV-immortalized human B cells, a potential natural target for c-MYC activation in EBV-positive AIDSNHL, leads to their malignant transformation (Lombardi et al., 1987). In addition, antisense oligonucleotides directed against translocated c-MYC mRNA are able to revert the tumorigenicity of lymphoma cells carrying an activated c-MYC (McManaway et al., 1990).In vivo, the targeted expression of c-MYC oncogenes in the B-cell lineage of transgenic mice leads to the development of B-cell malignancy at a relatively high frequency (Adams et al., 1985). In addition, it has been shown that c-MYC oncogene activation, besides deregulating B-cell growth and differentiation, may also cause other cellular alterations that may be positively selected in the context of immunodeficiency (Billaud et al., 1990; Inghirami et al., 1990; Klein, 1989b, 1994).One of these alterations may be the down-regulation of the integrin receptor lymphocyte function antigen-I (LFA-l), which is controlled by c-MYC in B cells and which is involved in a variety of biological functions, including cell conjugate formation between B cells and cytotoxic T cells, natural killer cells, and vascular endothelia (Inghirami et al., 1990). The c-MYC-induced downregulation of LFA-1 may be of particular relevance in the instance of EBV-infected B cells, since their recognition by cytotoxic T cells is mediated by the LFA molecules (Billaud et al., 1990; Gregory et al., 1988). Therefore, the occurrence of a c-MYC rearrangement in a B cell already carrying EBV will have at least two distinct effects. On one side, it will increase growth and confer full tumorigenic potential. On the other side, it will down-regulate the surface LFA antigens and thus allow tumor cells to escape immunosurveillance control. Our present insights into the biological role of c-MYC in B-cell lymphomagenesis are likely to be revolutionized by the identification of proteins that behave as c-MYC functional partners and form heterodimeric complexes with c-MYC (Blackwood and Eisenman, 1991; Gu et a.l., 1993, 1994). In the case of one of these, termed M a , it is already

136

GIANLUCA GAIDANO A N D RICCARDO DALLA-FAVERA

apparent that c-i\ZYC and Max together, but not c-MYC alone, can directly regulate gene transcription; furthermore, differential target gene expression and growth phenotype can be obtained by varying the c-MYCIMax ratio, with the heterodimeric c-MYC/Max complex acting as a positive regulator of transcription and growth, while Max alone is a negative regulator (Gu et al., 1993). Therefore, in AIDS-SNCCL or other lymphoid tumors carrying c-MYC translocations, it may be postulated that constitutive expression of c-MYC induces increased formation of c-MYCIMax complexes leading to positive growth stimulus. 2. BCL-6

Besides activating c-MYC, AIDS-NHL chromosomal translocations involve the BCL-6 gene (Table 111; Gaidano et al., 1994). T h e BCL-6 gene is a novel candidate protooncogene belonging to the family of transcription factors containing zinc finger domains (Baron et al., 1993; Kerckaert et ul., 1993; Ye et ai., 1993a,b).BCL-6 maps to 3q27, the site of frequent chromosomal breaks in B-cell NHL of the immunocompetent host and also occasionally involved in AIDS-NHL (Bastard et al., 1992; Bastard and 'l.illy, 1993; Offit et al., 1989). Chromosomal translocations between 3q27 and a heterogeneous chromosomal partner cause the 5' truncation of BCL-6 in more than 40% of DLCL of the immunocompet-ent host (Ye rt of., 1993b). T h e association between DLCL of the immunocompetent host and BCL-6 is selective, since BCL-6 rearrangements usually are not found in other NHL types (Lo Coco et ul., 1994). In addition, the finding of a BCL-6 rearrangement harbors clinical implications, documented by the fact that BCL-6 is a favorable prognostic marker and displays preferential clustering with extranodal NHL (Offit et al., 1994). As expected, BCL-6 rearrangements among AIDS-NHL are confined to AIDS-DLCL (20%), while they are consistently absent from AIDS-SNCCL (Gaidano et al., 1994). Only systemic AIDS-DLCL have been tested, and future studies will verify whether BCL-6 is also implied in AIDS-YCNSI,, which Lisually display DLCL morphology. Among AIDSDLCL, BCL-ti rearrangements are detected in both EBV-positive and EBV-negative samples and in both AIDS-DLCL variants, i.e., AIDSLNCCL and AIDS-LC-IBPL (Gaidano et al., 1994). Curiously, on the basis of the limited number of cases tested, BCL-6 truncations and ('-iVII'Cacti\.ation, although both present individually in a similar fraction of AIDS-DLCL (20%.),are never found in the same biopsy, suggesting that these two genetic lesions represent mutually exclusive molecular pathways in AIDS-DLCL development (Gaidano et al., 1994; Gaidano and Dalla-Favera, 1994a).

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

137

3. Other Loci Other dominantly acting oncogenes commonly involved in the pathogenesis of lymphomas of the immunocompetent host (e.g., BCL-1, BCL-8) do not seem to be relevant in AIDS-associated lymphomagenesis (Table 111; Gaidano and Dalla-Favera, 1992a; Gaidano et al., 1993; Subar et al., 1988). The consistent absence of BCL-2 rearrangements from AIDSDLCL testifies that these lymphomas are not preceded by a follicular phase, which instead is observed in a fraction of DLCL of the immunocompetent host (Lo COCOet al., 1993), but rather originate de nouu (Dalla-Favera et al., 1993). Finally, the RAS family of genes is known to be involved in AIDSrelated lymphomagenesis (Table 111; Ballerini et al., 1992, 1993). The mechanism of RAS activation reflects that commonly observed in human cancers, i.e., amino acid substitutions clustered at codons 12, 13, and 61 (Gaidano and Dalla-Favera, 1992b). Although the fraction of AIDSNHL cases carrying an activated RAS gene is low (15%),it is intriguing since NHL of similar histology arising in the immunocompetent host are consistently devoid of RAS mutations (Neri et al., 1988b).The biological significance of the selective association of RAS activation with AIDSNHL remains obscure. It is likely that mutated RAS genes, where present, might actively contribute to AlDS-NHL pathogenesis, since their role in the tumorigenic conversion of EBV-infected B cells is well established in vitro (Seremetis et al., 1989). B. TUMOR SUPPRESSOR LOCI

Few tumor suppressor loci have been investigated in AIDS-NHL. At present, evidence for a role of tumor suppressor genes in AIDS-related lymphomagenesis is limited to the case of p53 and deletions of the long arm of chromosome 6. 1 . p53

The role of the p53 gene has become evident from a number of observations that have documented its inactivation in 60% of AIDSSNCCL (Table 111) (Ballerini et al., 1992, 1993; Gaidano et al., 1993).In a large series from multiple institutions in the United States and in Europe, no p53 mutations were detected in AIDS-DLCL both by mutational analyses and by immunohistochemistry approaches (Ballerini et al., 1992, 1993; Gaidano et al., 1994; Carbone et al., 1994). In this respect, AIDS-NHL mirror the pattern ofp53 inactivation in NHL of the immunocompetent host arising de novo, in which $53 lesions are restricted to

138

GIAKLUCA G A I D A N O A N D RICCARDO DALLA-FAVERA

BL and never detected in de nozio DLCL (Gaidano et al., 1991). T h e lack of both BCL-2 rearrangements and p53 mutations in AIDS-DLCL suggests that AIDS-DLCL arises de no710 and is not preceded by a follicular phase (Lo Coco et al., 1993). T h e mechanisms of inactivation ofp53 in AIDS-NHL are identical to the ones occurring in other types of cancer and involve biallelic inactivation through point mutation of one allele and chromosomal deletion of the other (Hollstein et al., 1991). T h e mutational spectrum o f p 5 3 is highly heterogeneous and depends upon the mutagenic action of chemical oI physical carcinogens (Hollstein et al., 1991). I n the case of AIDSNHL, as also observed in other lymphoid tumors, brain tumors, and colon cancers, the mutations most frequently encountered are transitions at CpG dinucleotides (Ballerini et ai., 1993; Gaidano et al., 1991). This type of mutation is thought to occur as a DNA replication error, with no direct causal relationship with any known carcinogen (Hollstein et al., 1991). Mutations of p53 in AIDS-SNCCL occur in both EBV-positive and EBV-negative samples, suggesting that there is no direct correlation between these two genetic lesions (Ballerini el a/., 1983; Gaidano et al., 1993). Conversely, the frequent association between c-MYC deregulation and p53 inactivation in AIDS-SNCCL (Ballerini et al., 1993), as well as in BL of the immunocompetent host (Gaidano et al., 1991), may underly a synergistic effect of these two genetic lesions in the tumors. Finally, the biological role of p53 inactivation in AIDS-SNCCL is supported by the evidence that p53 wild-type genes are able to suppress the tumor growth of SNCCL cell lines carrying inactivating lesions of the cellularp53 genes (our unpublished observation). 2. Other Loci

Deletions of the long arm of chromosome 6 appear on an additional site of a putative tumor suppressor gene relevant to AIDS-related lymphomagenesis, being present in a substantial proportion (25%) of AIDSNHL (our unpublished observation). Deletions of 6q have been long since identified as one of the predominant genetic lesions of B-cell NHL, of which they represent a poor prognostic indicator (Offit et ul., 1991). Molecular mapping studies indicate that 6q deletions of B-cell NHL cluster in two discrete regions of minimal deletion along the long arm of chromosome 6 mapping to 6q27 (region of minimal deletion-1, RMD-1) and 6q21-q23 (RMD-2) (Gaidano et al., 1992, 1995).These two regions represent the site of two putative tumor suppressor genes relevant to lymphomagenesis (Gaidano et al., 1992, 1995). Interestingly, these same chromosomal regions have also been identified as the relevant

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

139

RMD of 6q deletions displayed by other tumor types, including acute lymphoblastic leukemia, malignant melanoma, and renal cell and ovarian carcinomas (Hatashi et al., 1990; Millikin et al., 1991; Morita et al., 1991). Fine genetic mapping of these deletions, as well as thorough hunting of the relevant tumor suppressor gene(s) is the field of active investigations. Among B-cell NHL, RMD-1 and RMD-2 display a preferential association with low and high grade lymphomas, respectively (Offit et al., 1993). It is thus expected that the RMD-2 lesion is the most critical in AIDS-NHL pathogenesis. The contribution of other tumor suppressor genes to AIDS-NHL pathogenesis has been ruled out, as in the case of RBI (Ballerini et al., 1993), o r awaits to be tested (Table 111).A putative involvement of NF- 1, APC, DCC, o r WT-1 seems unlikely on the basis of the lack of cytogenetic aberrations at the corresponding chromosomal sites in AIDS-NHL. On th,e other hand, the frequency of 9p deletions in high grade B-cell NHL (Chaganti et al., 1995) prompts a thorough search of structural lesions affecting the p16 tumor suppressor gene in AIDS-NHL (Kamb et al., 1994). VII. Conclusions: Distinct Pathogenetic Pathways in the Development of AIDS-Related Lymphomas

AIDS-NHL is a strikingly heterogeneous disease. At least two major pathogenetic pathways can be identified in AIDS-related lymphomagene s k Each of these pathogenetic pathways associates with peculiar clinical features and is restricted to distinct AIDS-NHL histological types (Fig. 2). T h e first of these pathways associates with AIDS-SNCCL, the equivalent of Burkitt’s lymphoma in the general population, and is characterized by relatively mild immunodeficiency of the host and multiplicity of genetic lesions of the tumor. AIDS-SNCCL generally occurs in the early phases of HIV infection while CD4 counts are still sustained, and its development appears to be relatively independent of prolonged exposure to immune suppression (Beral et al., 1991; Boyle et al., 1990; R(oithmann et al., 1991). More frequently than other AIDS-NHL, AIDSSlVCCL is preceded by a PGL phase (Kalter et al., 1985); this fact, together with the antigenic specificity and the immunogenetic features of AIDS-SNCCL (Riboldi et al., 1994), indicates that antigen stimulation and selection play important roles in the pathogenesis of these tumors. At the molecular level, AIDS-SNCCL is characterized by deregulation of c-MYC, inactivation of fi53,deletion of 6q, and mutation of RAS in 100, 60, 20, and 15% of the cases, respectively (Ballerini et al., 1993; Gaidano et al., 1993, 1994). Infection by EBV is restricted to 30% of AIDS-

In c

v) .-

h

m

b

s , I h

8 N

1.-

I

I

I

2 0 :

s

$ 3

I I

f I

m ln

I

8

I I I I I I

I I I

I

I I

I

I

(? 0

1

I

, ,

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

141

SNCCL, and its pathogenetic role is obscure (Ballerini et al., 1993; Gaidano et al., 1993, 1994). Typically, EBV-infected AIDS-SNCCL fail to express the viral transforming antigens EBNA-2 and LMP- 1 (Carbone et al., 1993b; Hamilton-Dutoit et al., 1993). In contrast to BL of the immunocompetent host, AIDS-SNCCL occasionally displays atypical morphological features resembling those of immunoblasts, particularly in EBVpositive cases associated with profound immunodeficiency. It is assumed that host immune regulation modulates AIDS-SNCCL phenotype, causing an apparent shift to immunoblastic-like lesions as host CD4 counts decrease (Delecluse et al., 1993). The most puzzling issue regarding AIDS-SNCCL is its unique association with AIDS among immunodeficiencies in humans. Unraveling of this discrepancy, putatively due to a causative agent closely linked to HIV and absent from other forms of immunodeficiency, may well enhance our understanding of the role of viral agents and the immune system in the development and control of cancer. A second major pathogenetic pathway is associated with the AIDSDLCL variants AIDS-LNCCL and AIDS-LC-IBPL, as well as with AIDSPCNSL. These tumors are characterized by severe immunodeficiency of the host and a very poor prognosis (Beral et al., 1991). The actual role of prolonged exposure to immunosuppression in the genesis of AIDSDLCL and AIDS-PCNSL is testified by the outburst of these neoplasms in long-term AIDS survivors and by the direct correlation between decreased CD4 counts and increased hazards of lymphoma development

FIG.2. Clinicopathologic heterogeneity of AIDS-NHL. AIDS-SNCCL tends to arise as a relatively early event after HIV infection and is molecularly characterized by c-MYC rearrangements, pS3 inactivation, and EBV infection in 100, 60, and 30% of cases, respectively. R A S mutations and 6q deletions, although not specific for AIDS-SNCCL, may also occur. In contrasi LO AIDS-DLCL and AIDS-PCNSL, EBV-infected AIDS-SNCCL d o not express the viral transforming antigens EBNA-2 and LMP-I. AIDS-SNCCL may develop in the presence of relatively well preserved host immunosurveillance. as demonstrated by peripheral blood CD4 counts, and harbors a relatively favorable prognosis when compared to other AIDS-NHL types. AIDS-DLCL generally develops as a late manifestation of HIV infection, is strictly dependent upon lowering of peripheral blood CD4 counts, and is characterized by a poor prognosis. Infection by EBV, generally expressing EBNA-2 and LMP-I, is the most frequent genetic lesion, whereas rearrangements of BCL-6 and c-MYC are restricted to a fraction of cases. AIDS-PCNSL mimicks many clinical and biological features of AIDS-DLCL, including late presentation after HIV infection and low CD4 counts. AIDS-PCNSL harbors the worst prognosis among AIDS-NHL. AIDS-PCNSL pathogenesis is characterized by EBV infection in virtually all cases. In addition to EBNA-2 and LMP-1, rare AIDS-PCNSL cells also consistently express the EBV lytic cycle protein VCA.

142

GIAKLUCA GAIDANO AND RICCARDO DALLA-FAVERA

(Karp and Broder, 1991).Traditionally, AIDS-DLCL and AIDS-PCNSL have been thought of as EBV-driven lymphoproliferations arising in the context of a highly disrupted cytotoxic control directed against EBV. This view is supported by the apparent lack of genetic lesions other than EBV infection in the overwhelming majority of cases (Ballerini et al., 1993; Gaidano et al., 1994). Since both EBNA-2 and LMP-1 are expressed in EBV-positive AIDS-DLCL and in AIDS-PCNSL (Carbone et al., 1993b; Haniilton-Dutoit et al., 1993), it is reasonable to think that EBV is indeed a driving force for tumor growth and expansion. With respect to these pathogenetic features, AIDS-DLCL and AIDS-PCNSL would be similar to the lymphomas encountered in association with congenital or iatrogenic immunosuppression and the Iymphoproliferations arising in animal models after inoculation of EBV-infected B cells (Locker and Nalesnick, 1989). T w o aspects, however, differentiate AIDSDIXL and AIDS-PCKSL from lymphomas associated with congenital or iatrogenic immunosuppression. First, whereas the latter may regress upon recovery of the patient’s T-cell function (Locker and Nalesnick, 1989), rZIDS-related cases inexorably progress. Second, in contrast to lymphomas associated writh congenital or iatrogenic immunosuppression, and in contrast to the lymphoproliferations developing in animal models, AIDS-DLCL are consistently monoclonal. In rare cases, the monoclonality may be accounted for by the presence of genetic lesions of c-MYC and BCL-6, although in most cases no genetic lesions are detected. I t is possible that the AIDS-DLCL phenotype may be caused by a variety of genetic pathways, which have in common EBV infection and profound host immunodepression, but differ in the nature of genetic lesions contributing to the lymphoma. Although many aspects of AIDS-related lymphomagenesis remain unclear, the clinicopathologic and experimental investigations performed in the AIDS-KHL field during the last decade have unexpectedly enriched our knowledge of how and why NHL develop. Some general concepts of human neoplasia have also been gained from the model of AIDS-related lymphomagenesis. In particular, the striking clinicopathologic heterogeneity of AIDS-NHL has instructed us on how the action of the same risk factor (HIV-induced immunodeficiency in the case of AIDS) on the same tissue (B cells in the case of AIDS) may be responsible for substantially different tumors that are sustained by different genetic lesions and display different clinicobiologic behaviors. In all of these respects, AIDS-related lymphomagenesis represents a unique system to study the contribution of host-tumor interactions in the development of human neoplasia.

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

143

ACKNOWLEDGMENTS Work described in this chapter has been supported by NIH Grant CA-37295 and by VIII AIDS Project (Grant No. 9206-ll), ISS, Rome. The authors are grateful to several collaborators who have contributed many of the results described in this review, particularly Drs. D. M. Knowles, A. Neri, M. Subar, P. Ballerini, and P. Casali.

REFERENCES

Aclams, J. M., Harris, A. W., Pinkert, C. A., Corcoran, L. M., Alexander, W. S., Cory, S., Palmiter, R. D., and Brinster, R. L. (1985). Nature 318, 533-538. Agnarsson, B. A.. and Kadin, M. E . (1988). Am.]. Surg. Pathol. 14, 264-274. Ahrned, T., Wormer, G. P., Stahl, R. E.,Mamtani, R., Cimino, J., Glasser, R., Mittelman, A., Friedland, M., and Arlin, 2. (1987). Cancer 60, 719-723. Amadori, A., Zamarchi, R., Veronese, M. L., Panozzo, M., Barelli, A., Borri, A,, Sironi, M., Colotta, F., Mantovani, A,, and Chieco-Bianchi, L. (1991).J Immunol. 146, 57-62. Amariglio, N., Vonsover, A,, Hakirn, I., Neumann, Y.,Mark, Z., Brok-Sirnoni, F., BenBassat, I., and Rechavi, G. (1994). Acta H m t o l . 91, 103-105. A yers, K. M. (1988). Am. ]. Med. 85, 186- 188. Ballerini, P., Gaidano, G., Gong, J. Z., Tassi, V., Saglio, G., Knowles, D. M., and DallaFavera, R. (1992). AIDS Res. Hum. Retrouinues 8, 731-735. Ballerini, P., Gaidano, G., Gong, J. Z., Tassi, V., Saglio, G., Knowles, D. M., and DallaFavera, R. (1993). Blood 81, 166-176. Baron, B. W., Nucifora, G., McCabe, N., Espinosa, R., 111, Le Beau, M. M., and McKeithan, T. W. (1993). Proc. Natl. Acad. Sci. USA 90, 5262-5266. Bashir, R. M., Harris, N. L., Hochberg, F. H., and Singer, R. M. (1989). Neurology 39,813817. Bashir, R., Luka, J., Cheloha, K., Chamberlain, M., and Hochberg, F. (1993).Neurology43, 2358-2362. Bastard, C., and Tilly, H. (1993). Blood 81, 265 (letter). Bastard, C., Tilly, H., Lenorrnand, B., Bigorgne, C., Boulet, D., Kunlin, A., Monconduit, M., and Piguet, H. (1992). Blood 79, 2527-2531. Baumgartner, J. E., Rachlin, J. R., Beckstaed, J. H., Meeker, T. C., Levy, R. M., Wara, W. M., and Rosenblum, M. L. (199O).J. Ncurosurg. 73, 206-211. Benjamin, D., Knobloch, T. J., and Dayton, M. A. (1992). Blood 80, 1289-1298. Beral, V., Peterman, T., Berkelman, R., and Jaffe, H. (1991). Lancet 337, 805-809. Berard, C. W., OConnor, G. T., Thomas, L. B., and Torloni, H. (1989).Bull. World Health Organ. 40,601-607. Brrek, C., and Milstein, C. (1987). Immunol. Rev. 96, 23. Bernheim, A,, and Berger, R. (1988). Cancer Genet. Cytogenet. 32, 67-74. Br:rnstein, L., and Hamilton, A. S. (1993). Cum. Opin. Oncol. 5, 822-830. Bernstein, L., Levin, D., Menck, H., and Ross, R. K. (1989). Cancer Res. 49, 466-470. Bhatia, K., Huppi, K., Spangler, G., Iyer, R., and Magrath, I. (1993). Nature Genet. 5 , 56-6 1. Bhatia, K., Spangler, G., Gaidano, G., Hamdy, N., Dalla-Favera, R., and Magrath, I. (1994). Blood 84,883-888. Biggar, R. J., and Rabkin, C. S. (1992). Cum Opin. Oncol. 4, 883-893.

144

GIANLUCA GAIDANO AND RICCARDO DALLA-FAVERA

Bahler, D. W., and Levy, R. (1992).Proc. Natl. Acad. Sci. USA 89,6770-6774. Billaud, M., Rousset, F., Calender, A., Cordier, M., Aubry, J. P., Laisse, V., and Lenoir, G. hI. (1990).Blood 75, 1827-1833. Birx, D., Redfield, R., and Tosato, G. (1986).hime Engl. J. Med. 314,874-879. Birx, D. L., Redfield, R. R., Tencer, K., Fowler, A., Burke, D. S., and Tosato, G. (1990). Blood 76,2303-23 10. B l a c k w d . E. M.. and Eisenman. R. N. (1991).Science 251, 121 1-1217. Borisch, B., Ellinger. K., Neipel, F.. Fleckenstein, B., Kirchner, T., Ott, M. M., and MullerHernielinck, H. K. (1991).Virch. Arch. B CellPathol. 61, 179-187. Boyle, hf. J., Swanson, C. E., Turner, J. J., Thompson, I. L., Roberts, I., Penny, R., and Cooper, D. A. (1990). Br.J. Haematol. 76, 506-512. Breen, E. C., Rezai. A. R., Nakajima, K., Beall, G. L., Mitsuyasu, R. T., Hirano, T., Kishimoto, T., and Martinez-Maza. T. (1990).J.Immunol. 144,480-484. Brown, N. A., Liu, C. R., Wang, Y. F., and Garcia, C. R. (1988).J.Virol. 62, 962-969. Carbone, A.,Tirelli, U.,Vaccher, E., Volpe, R., Gloghini, A., Bertola, G., De Re, V., Rossi, C., Boiocchi, M.,and Monfardini. hi. (1991).Ca.ncer68,842-852. Carbone, A., Gloghini, A., Zanette, I., Canal. B., and Volpe,R. (199.3,). Am. J. Clitz. Pathol.

99,289-297.

Carbone, A., Tirelli, U.,Gloghini, A., Volpe, R., and Boiocchi, M.(1993b).J.Cltn. O t ~ o l . 11, 1674-1681. Carbone, A., Gloghini, A., Gaidano, G., Cilia, A. M., Bassi, P., Polito, P., Vaccher, E.,Saglio, G., and Tirelli, U. (1995).A n . /. Cltn. Pathol. 103,561-567. Casabona, J., Melbye, M.. Bigar, R., and The AIDS Registry Contributors (1991).Int. J. Cancer 47,49-53. Centers for Disease Control (1985).it4MW’R Morb. Mortal Wkly. Rep. 34,373-375. Cesarman, E., Dalla-Favera, R., Bentley, D., and Groudine, M. (1987).ScWnce 238, 1272-

1275.

Chadburn. A.,Cesarman. E., Jagirdar, J.. Subar, M..Mir, R. N., and Knowles, D. M. (1993). Canrrr 72,3078-3090. Chaganti, R. S. K.. Jhanwar, S. C., Koziner. B., Arlin, Z., Mertelsmann, R., and Clarkson, B. C. (1983).Blood 61, 1269-1272. Chaganti, S., Gaidano, G.. Dalla-Favera, R., and Chaganti, R. S. K. (1995).Genes,C h r m some.v, Cancer. 12,32-36. Cinque, P.,Brytting. M.,Vago, L., Castagna, A.. Parrwicini, C.. Zanchetta, N., DArminio Monforte, .4.. Wahren, B., Lazzarin, A.. and Linde. A. (1993).Lancet 342, 398-401. Cohen, J. I., Wang, F.. Mannick, J., and Kieff, E. (1989).Proc. h’atl. Acad. Sn’. USA 86,

9558-9562.

Contreras-Brodin. R. A.. Anvret, hl.. Imreh, S.,Altiok, E., Klein. G., and Masucci, M. G. (1991).J. Grn. V i r d 72, 3025-3033. Cremer, K. J., Spring, S. R., and Gruber, J. (199O).J.,\’at/. Cancm f w t . 82, 1016-1024. Dalla-Favera. R. ( 1‘393). Chromosomal translocations involving the c-MYC oncogene in lymphoid nmplasia. I n ‘The Causes and Consequences of Chromosomal Aberrations” (I. R. Kirsch. Ed.). pp. 313-332. CRC Press, Boca Raton, FL. Dalla-Favera, R., Bregni, M.,Erickson, J.. Patterson, D. Gallo, R. C. and Croce, C. M. ( 1981).Prw. ,Vat/. Acad. Sci. USA 79, 7824-7827. Dalla-Favera. R., Martinotti, S., Gallo. R. C., Erickson, J.. and Croce, C. XI. (1983).SciPtue

219,963-967.

Dalla-Favera. R.. Ye. B. H.. Lo Coca, F., Gaidano. G., Lista, F., Knowles, D. M.,Louie, D. C., Offit, K., and Chaganti, R. S. K. (1993).Ann. Oncol. 5, S55-S60. Davis. hi., Xlalcolm. S., and Rabbits, T. 11. (1984).Snturr 308, 286-288.

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

145

De Angelis, L. M., Wong, E., Rosenblum, M., and Furneaux, H. (1992). Cancer 70, 16071611. Del’ecluse, H. J., Raphael, M., Magaud, J. P., Felman, P., Abd Alsamad, I., Bornkmann, (G. W., Lenoir, G. M., and the French Study Group of Pathology for Human Immuinodeficiency Virus-Associated Tumors (1993). Blood 82, 552-563. de IMiranda, P., Burnette, T. C., and Good, S. S. (1990). DrugMetab. Dispos. 18, 315-320. de Waal Malefyt, R., Haanen, J., Spits, H., Roncarolo, M.-G., te Velde, A., Figdor, C., Johnson, K., Kastelein, R., Yssel, H., and de Vries, J. E. (1991).J. Exp. Med. 174, 915924. Di Carlo, E., Amberson, J., Metroka, C., Ballard, P., Moore, A., and Mouradian, J. (1986). .4rch. Pathol. Lab. M i d . 10, 1012-1016. Doll, D. C., and List, A. F. (1982). Lancet i, 1026-1027. Eick, D., and Bornkamm, G. W. (1989). E M B O J . 8 , 1965-1972. Emilie, D., Peuchmaur, M., Maillot, M. C., Crevon, M. C., Brousse, N., Delfraissy, J. F., .Dormant, J., and Galanaud, P. (199O).J. Clzn. Invest. 86, 148-159. Emilie, D., Coumbaras, J., Raphael, M., Devergne, O., Delecluse, H. J., Gisselbrecht, C., Michiels,J. F., Van Damme, J., Taga, T., Kishimoto, T., Crevon, M. C., and Galanaud, P. I[ l992a). Blood 80, 498-504. Emilie, D., Touitou, R., Raphael, M., Peuchmar, M., Devergnee, O., Rea, D., Coumbaras, J., (Crevon, M. C., Edelman, L., Joab, I., et al. (1992b). Eur. J. Immunol. 22, 2937-2942. Epstein, M. A., Achong, B. M., and Barr, Y. M. (1964). Lancet 1, 702-703. Fauci, A. S., Schnittman, S. M., Poli, G., Koenig, S., and Pantaleo, G. (1991). Ann. Intern. Med. 114, 678-693. Feichtinger, H., Putkonen, P., Parravicini, C., Li, S.-L., Kaaya, E. E., Bottiger, D., Biberfeld, G., and Biberfeld, P. (1990). A m . ] . Pathol. 137, 1311-1315. Fiorentino, D. F., Zlotnik, A,, Vieira, P., Mossmann, T. R., Howard, M., Moore, K. W., and {OGarra,A. (1991).J . Immunol. 146, 3444-345 1. Fischl, M., Richman, D. D., Grieco, M. H., Gottlieb, M. S., Volderbing, P. A., Laskin, 0. L., Leedom, J. M., Groupman, J. E., Mildvan, D., Schooley, R. T., Jackson, G. G., Durack, D. T., and King, D. (1987). New Eng1.J. Med. 317, 185-191. Formenti, S. C., Gill, P. S., Lean, E., Rarick, M., Meyer, P., Boswell, W., Petrovich, Z., Chak, L., and Levine, A. M. (1989). Cancer 63, 1101-1107. Friedman, D. F., Cho, E. A,, Goldman, J., Carmack, C. E., Besa, E. C., Hardy, R. R., and ,SiIberstein, L. E. (1991).J. Exp. Med. 174, 525-537. Frizzera, G., Rosai, J., Dehner, L. P., Spector, B. A,, and Kersey, J. H. (1980). Cancer 46, 692-699. Gai,dano, G., and Dalla-Favera, R. (1992a). Curr. Opzn. Oncol. 4, 900-906. Gai,dano, G., and Dalla-Favera, R. (199213). Proto-oncogenes and tumor suppressor genes. In “Neoplastic Hematopathology” (D. M. Knowles, Ed.), pp. 245-261. Williams & Wilkins, Baltimore. Gaitdano, G., and Dalla-Favera, R. (1993). Curr. Opin. Oncol. 5 , 776-784. Gaitdano, G., and Dalla-Favera, R. (1994). Antzbzot. Chemother. 46, 117-124. Gabdano, G., and Dalla-Favera, R. (1995). Molecular biology of lymphoid neoplasms. In “The Molecular Basis of Cancer” (J. Mendelsohn, P. M. Howley, M. A. Israel, and L. A. Liotta, Eds.), pp. 25 1-272. WB Saunders, Philadelphia. Gaidano, G., Ballerini, P., Gong, J. Z., Inghirami, G., Neri, A,, Newcomb, E. W., Magrath, [. T., Knowles, D. M., and Dalla-Favera, R. (1991). Proc. Natl. Acad. Scz. USA 88, 54135417. Gaitdano, G., Hauptschein, R. S., Parsa, N. Z., Offit, K., Rao, P. H., Lenoir, G. Knowles, D., Chaganti, R. S. K., and Dalla-Favera, R. (1992). Blood 80, 1781-1787.

146

LIAICLUCA GAIDAN'O A N D RICCARDO DALLA-FAVERA

Gaidano. G . , Parsa, S . 2.. Tassi. V., Della-Latta, P., Chaganti. K.S. K., Knowles, D. M., and Dab-Favera, K. ( 1993)Lrukevrrn 7, 162 1 - 1629. Gairiano. (;,, Lo (;(KO, F., Ye, I).H., Shibata. D., Levine, A . M., Knowles, D. M., and DallaFavera. R. ( 1994a). Blood 84, 397-402. Gaidano, G., Hauptschein, R. S., Venkatraj, V. S., Rutner, T., Rao, P. H., Chaganti, R. S. K., and Dalla-Favera, R. (1994b). 6q deletions in B-cell non-Hodgkin lymphoma. In "Molecular Diagnosis and Monitoring of Leukaemia and Lymphoma, Challenges of Mode r n Medicine" (E. C. Borden, J. M., Goldman, and F. Grignani, Eds.), Vol. 2, pp. 9-19. Ares Serono Press, Rome. Gail, M. H.. Plrtda. J. M.,Rahkin, C. S., Biggar, R. J., Goedert, J. J., Horm, J. W., Sondik, E. J., Yarchoan, R., and Broder. S. (l991).J. 5d.Cnnrer I m t . 83, 695-701. Ganser. A , , Carlo-Stella, C., Bartram, C. K . . Boehni, Th., Heil, G., Hengkin, B., Muller, H., Kagha\achar. An., v. Briesen. H . , Griesinger, F., Volkers, B., RuebsamenHelm, E. B., and Hoelzer, D. (1988). Blood 72, 1259-1260. IVaigrnann, H., and GOOC~. R. A. (1971). Cnrrcer 28, 89-98. Gatti. K. .I.. Gill. P. S.. Levine. .A. M., Meyer. P. K.. BOSWKII, W.D., Burkes, R. L., Parker, J. W., Hofrnan, F. SI.. D\voIsk\, K. L., a n d Lukes. K.J. (198.5). 29~n.].Mrd. 78, 742-748. (;o. B . N . F.,Castle. B. E., Barret, K., Kastelein, K.. Dang, W., Mossniann. '1.. K., Moore, K. W., and Howard, %I. (1990).J. Exp. .\fed. 172, 1625-1631. Goldstein, J. D., Dicksoti, D. I V . , \loser, F. G . , Hirschfeld, A . D., Freeman, K., Leena, J., Kaplan, D.. and Davis, L. (1991). Cancet 67, 2756-2765. Gregory. C. D.. Xlurray. R. J.. Edwards, C. F., and Rickinson, A. B. (1988).J.Ex/), M e d . 167, 1811-1824. Grignani. F.. Lombartii. L.. Inghiranii. G.. Sternas, L.. Cechova. K.,Knowles, D. hf., and Dalla-Faveta, R. (1990). E J f B O J . 9, 3915-9922. (;roopinan, J. E., Sullivan. J . L.. hlulder. C . (;insburg, ~ D., Orkin, S . H., O'Hara, C. J,, Falchuk, I(..TVOtig-Stad. F., 2nd Gallo. R. C. (1986).Blood 67, 612-615. (;ti, b'.,Ccchova, K.. ?'xi,\'., and Dalla-Favei-a. R. ( 1993). !'roc. Nntl. Arad. Scz. USA 90, 2935-2939. (hi. W.. Bhatia. K., Sfagrath. I . T.. Dang, C. N'.,and Dalh-Fdvera, R. (1994). Science 264, 2 3 1-254. (;uarner. .I.. rlel Rio. C. Carr, D.. Hentlris, I.. E., Ele); J. M:, and Unger, E. R. (1991). C n i i r r t 68, 2160-246.5. Haluska. F. (;.. Finger. S.. l'sujimoto, Y.,and Croce, C. hf. (1986). "Vuture 324, 158-161. Haluska, F. (;., Tsujirnoto, Y., and Croce, C. hl. (1987). A r t ! i i ~Re11. Genet. 21, 321-345. Ilaluska. F. (;., Kusso. (;., Kant, J., Andreef, Sl., and Croce, C:. M.(1989).Pror. h'crtl. Acad. Sri. L.S.9 86, X W - 8 9 1 1 . Ilamiltoii-Dutoit, S. J.. Pallesen, G., Franzniann. M. B.. Karkov, ,J., Skinhflj, P., and P K d rrsen. P. (1991). ..iin.J. I-'c2/hol. 138, 149-163. Hamilton-Dutoit. S.J.. Kaphael. 11.. Audouin, .J., Diebold, .J.. Lisse, I . , Pedersen, C., Oksenhendler. E., Slarelle, L., and Pallesen, G. (1993). Blood 82, 619-624. Harnl). 31. E., % a i . S. H.. Holly, E. A , , Kelter. .I.. and Padian, N.(1988). A m , ] . Epidemiol. 128, 26 1-267. Harris. N . I_.,_jaffe. E. S., Stein. H.. Banks, P. M.,Chan, J. K. C., Cleary, M.,Delsol, G., De ?%'oIf-Peters.C.. Falini. 13.. (hirer. K. C;.. Grogan, -I:, Isaacson, P. C., Knowles, D. M., Mason. D. Y.,Sluller-Hermelink. H.-K.. Pileri, S. A.. Piris. M . A,, Ralfkiaer, E., and kVarmke. R. A . (1994). Blood 84, 1:161-1392. flatashi, Y..Kaiinondi. S.C., Look, A. T., Behni, F. G., Kitchingiiian, G. R . , Pui, <:.-H., Rivera, (;. I(..and It'illianis, D. 1.. (1990). Blood 76, 1626-1630.

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

147

Henderson, S., Rowe, M., Gregory, C., Croom-Carter, D., Wang, F., Longnecker, R., Kieff, E., and Rickinson, A. (1991). Cell 65, 1107-1 115. Herndier, B. G., Shiramizu, B. T., Jewett, N. E., Aldape, K. D., Reyes, G. R., and McGrath, M. S. (1992). Blood 79, 1768-1774. Hessl, N. A., Katz, M. A., Liu, J. Y., Buchbinder, S. P., Rubino, C. J., and Holmberg, S. D. (1992). Ann. Intern. Med. 117, 309-311. Hollis, G. F., Mitchell, K. F., Battey, J., Potter, H., Taub, R., Lenoir, G. M., and Leder, P. (1984). Nature 307, 752-755. Hollstein, M., Sidransky, D., Vogelstein, B., and Harris, C. C. (1991). Science 253, 4953. Hoover, R., and Fraumeni, J. F., Jr. (1973). Lancet 2, 55-57. Hui, P. K., Feller, A. C., and Lennert, K. (1988). Histopathology 12, 127-143. Ikematsu, H., Harindranath, N., Ueki, Y., Notkins, A. L., and Casali, P. (1993).J.Immunol. 150, 1325-1337. Inghirami, G., Grignani, F., Sternas, L., Lombardi L., Knowles, D. M., and Dalla-Favera, R. (1990). Science 250, 682-686. Ioachim, H. L. (1992). Leukemia 6 (Suppl. 3), 30s-33s. Ioachim, H. L., Cooper, M. C., and Hellman, G. C. (1985). Cancer 56, 2831-2842. Ioachim, H. L., Dorsett, B., Cronin, W., Maya, M., and Wahl, S. (1991). Hum. Pathol. 22, 659-673. Irwin, D., and Kaplan, L. (1993). C u m Ofin. Oncol. 5, 852-860. Italian Cooperation Group for AIDS-related Tumors. (1988).J. Natl. Cancer Imt. 80, 855860. Jaffe, H. W., Choi, K., Thomas, P. A., et al. (1983). Ann. Intern. Med. 99, 145-151. Kalter, S., Riggs, S., Cabanillas, F., Butler, J., Hagemeister, F., Mansell, P., Newell, G., Velasques, V., Salvador, P., Barlogie, B., Rios, A., and Hersh, E. (1985). Blood 66,655659. Karnb, A., Gruis, N. A., Weaver-Feldhaus, J., Liu, Q., Harshman, K., Tavtigian, S. V., Stockert, E., Day, R. S., 111, Johnson, B. E., and Skolnick, M. H. (1994). Science 264, 436-440. Kiplan, L. D., Abrams, D. I., Feigl, E., McGrath, M., Kahn, J., Neville, P., Ziegler, J., and Volderbing, P. A. (l989).J. Am. Med. Assoc. 261, 719-724. Kirp, J. E., and Broder, S. (1991). Cancer Res. 51, 4743-4756. Katz, B. Z., Berkman, A. B., and Shapiro, E. D. (1992).J. Pediatr. 120, 228-232. Kawano, M., Hirano, T., Matsuda, T., et al. (1988) Nature 332, 83-85. Kerckaert, J.-P., Deweindt, C., Tilly, H., Quief, S., Lecocq, G., and Bastard, C. (1993) Nature Genet. 5, 66-70. Kersey, J. H., Spector, B. D., and Good, R. A. (1973). Int. J. Cancer 12, 333-347. Kieff, E., and Liebowitz, D. (1989). Oncogenesis by herpes viruses. In “Oncogenes and the Molecular Origins of Cancer” (R. A. Weinberg, Ed.), pp. 259-280. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Ki+imoto, T. (1989). Blood 74, 1-10. Klatt, E. C. (1988).J . Acquir. Immune Defic. Syndr. 1, 459-465. Klein, G. (1975). New Eng1.J. Med. 293, 1353-1357. Klein, G. (1989a). Cell 58, 5-8. Klein, G. (1989b). Genes, Chromosomes, Cancer 1, 3-8. Klein, G. (1992). Blood 80, 299-301. Klein, G. (1994). Antibiot. Chemother. 46, 110-1 16. Klein, G., and Klein, E. (1985). Nature 315, 190-195.

148

GIANLUCA GAIDANO AND RICCARDO DALLA-FAVERA

Knowles. D. M.,and Chadburn, A. (1992). In “Neoplastic Hematopathology” (D. M. Knowles, Ed.), pp. 773-835. Williams & Wilkins, Baltimore, MD. Knowles, D. M., Chaniulak, hl. D., Subar, M., Burke, J. S., Dugan, M.. Wernz, J., Slywotzky. C.,Pelicci, P.-G.,Dalla-Favera, R., and Raphael, B. (3988). Ann. I n t n . Mcd. 108,744753.

Kopelman, R. G., and Zolla-Pazner, S. (1988). Am. J. Med. 84,82-88. Lane. H.C.,Masur, H., Edgar, L. C., Whalen, G., Rook, A. H., and Fauci, A. S.(1983). New E q / . J . Med. 309, 453-458. Laurence.J., and Astrin, S. M. (1991). Proc. Natl. Acad. Sci. USA 88,7635-7639. Levine, A. M., Meyer, P. R., Begandp M. K., Parker, J. W.. Taylor, C. R., Irwin, L., and Lukes, R. J. (1984). Ann. h t m z . Med. 100, 7-13. Levine, A. M., Sullivan-Halley, J., Pike, M. C., Rarick, M. U., Loureiro, C., BernsteinSinger, M., Willson, E.,Brynes, R., Parker, J., Rasheed, S.,and Gill, P. S. (1991). Cancer 68,2466-2472.

Levine, A. M. (1992). Blood 80,8-20. Levine, A. M. (1993). J. Natl. Cancer Insf. 85, 1382-1397. Levy, S., Mendel, E..Kon, S.,Avnur, Z., and Levy, R. (1988). J. Exp. Med. 168,475-489. Levy, Y.,Tsapis. A.. and Brouet, J. C. (1991).J. Clin. Invest. 88,696-699. List, A. F., Spier, C. M., Miller, T. P., and Grogan, T. M. (1993). Leukemia 7, 398-403. Locker, J.. and Nalesnick, M. (1989). Am.J. Patlwl. 135,977-987. Lo Coco. F., Gaidano, G., Louie, D. C., Offit, K., Chaganti, R. S. K., and Dalla-Favera, R. (1993). Blood 82,2289-2295. Lo Coco, F., Ye, B.H.. Lista. F., Corradini, P.. Offit, K., Knowks, D. M.. Chaganti, R. S.K., and Dalla-Favera, R. (1994). Blood 83, 1757-1759. Lomhrdt, 1..Xewcomb, E. W., and Dalla-Favera, R. (1987). CeU49, 161-170. Loureiro, C.. Gill. P.S.,Meyer, P. R., Rhodes, R., Rarick, M. U., and Levine, A. M. (1988). Cancer 62,739-739.

imwenthal, D. A. Strauss, D. J.. Campbell, S.W.,(;old, J. W., Clarkson, B. D., and Koziner, B. (1988). Cancer 61, 2325-2335. Lukes, R. 1.. Parker. J. W.,Taylor. C. R.. Tindle, B. H., Cramer, A. D., and Lincoln, T. 1. ( 1970). Semin. Hematol. 15, 322-33 1. MacMahan, E. M.E.,Class, J. D., Hayward. S.D., Mann, R. B., Becker, P. S.,Charache, P., McAdiur. J. C., and hmbinder, R. F. (1991). Lancet 338, 969-953. Magrath, I. T. (1990). Small non cleaved cell lymphomas. In “The Non-Hodgkh’s Lymphomas” (I. T. Magrath, Ed.), p p 256-278. Edward Arnold, London. Magrath, 1. T., Erikson, J.. \.\.‘hang-Peng,J.. Siverts, H.,Armstrong, G.,Benjamin, D., Triche, T.,Alabaster. O., and Croce. C. M. (1983). Science 222, 1094-1098. Martinez-Maza. 0.. Crabb, E., Mitsuyasu, R. T., Fahey, J. L., and Giorgi, J. V. (1987). J. Immunol. 138,3720-3724. Mathur-Wagh. U.,Enlorv, R. W.,Spigland, I., et at., (1984). Lancet i, 1033-1038. blchtanaway, M.E.,Neckers, L. M., Lake. S.L.. Al-Nasser, A. A., Redner, R. L.,Shiramizu, B. T., Coldxhmidts, W.L.. Huber, K. E., Bhatia, R.,and Magrath, I. T. (1990). Lancef 335,808-8 1 1.

Meeker, 1.C., Shiramizu, B., Raplan. L., Herndier, B.. Sanchez, H., Grimaldi, J. C., Baumgartner, J., Rachlin, J., Feigal, E., Rosenblum, hf., and McGrath. M. S. (1991). AIDS 5,669-674. Metroka, C.E.. CtinninRham-Rundles, S.. Pollack, ?ti.S., et al. (1983). .4nn. Ititen,. Med. 99, 385-59 I . Miller. B. A.. Ries. L. A. C., Hankey. B. F., Kosary. C. L., and Edwards, B. K. (1992).

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

149

“Cancer Statistics Review: 1973-1989,” NIH Publication Number 92-2789. National Cancer Institute, Bethesda, MD. Millikin, D., Meese, E., Vogelstein, B., Witkowski, C., and Trent, J. (1991). Cancer Res. 51, 5449-5453. Monfardini, S., Vaccher, E., FOP, R., Tire&, U., and Others for The Italian Cooperative Group on AIDS-Related Tumors ( 1 990). Ann. Oncol. 1, 203-2 1 1 . Moore, R. D., Kessler, H., Richman, D. D., Flexner, C., and Chaisson, R. E. (1991).JAMA 265,2208-22 11. Morgello, S. (1992). Am. J. Pathol. 141, 441-450. Morita, S., Saito, S., Ishikawa, J., Ogawa, O., Yoshida, O., Yamakawa, K., and Nakamura, Y. (1991). Cancer Res. 51, 5817-5820. Mloss, D. J., Misko, I. S., Burrows, S. R., Burman, K., McCarthy, R., and Sculley, T. B. (1988). Nature 331, 719-721. Murray, R.J., Wang, D., Young, L. S., Wang, F., Rowe, M., Kieff, E., and Rickinson, A. B. (1988).J . Virol. 62, 3747-3755. Nath, N., Wunderlich, C., Darr, F. W., 11, Douglas, D. K., and Dodd, R. Y. (1987).J. Clin. Microbiol. 25, 364. National Cancer Institute and Centers for Disease Control (1991).M M W R 40, 591, 597600. Neri, A,, Barriga, F., Knowles, D. M., Magrath, I., and Dalla-Favera, R. (1988a). Proc. Natl. Acad. Sci. USA 85,2748-2752. Neri, A., Knowles, D. M., Greco, A., McCormick, F., and Dalla-Favera, R. (1988b). Proc. Natl. Acad. Sci. USA 85, 9268-9272. Neri, A,, Barriga, F., Inghirami, G., Knowles, D. M., Neequaye, J., Magrath, I. T., and Dalla-Favera, R. (1991). Blood 77, 1092-1095. Ng, V. L., Hwang, K. M., Reyes, G. R., Kaplan, L. D., Khayam-Bashi, H., Hadley, W. K., and McGrath, M. S. (1988). Blood 71, 3197-3201. Ng, V. L., Chen, K. H., Kwang, K. M., Khayam-Bashi, H., and McGrath, M. S. (1989).Blood 74,2471-2475. Ng, V. L., Hurt, M. H., Fein, C. L., Khayam-Bashi, F., Marsh, J., Nunes, W. M., McPhaul, L. W., Feigal, E., Nelson, P., Herndier, B. G., Shiramizu, B., Ryes, G. R., Fry, K. E., and McGrath, M. S. (1994). Blood 83, 1067-1078. Non-Hodgkin’s Lymphoma Pathologic Classification Project (1982). Cancer 49,2 112-2 135. Olffit, K., Jhanwar, S., Ebrahim, S. A. D., Filippa, D., Clarkson, B. D., and Chaganti, R. S. K. (1989). Blood 74, 1876-1879. Olffit, K., Wong, G., Filippa, D. A,, Tao, Y., and Chaganti, R. S. K. (1991). Blood 77, 15081515. O’ffit, K., Parsa, N. Z., Gaidano, G., Filippa, D. A., Louie, D., Pan, D., Jhanwar, S. C., DallaFavera, R., and Chaganti, R. S. K. (1993). Blood 82, 2157-2162. Omffit, K., Lo COCO,F., Louie, D. C., Parsa, N. Z., Leung, D., Portlock, C., Ye, B. H., Lista, F., Filippa, D. A., Rosenbaum, A., Ladanyi, M., Jhanwar, S., Dalla-Favera, R., and Chaganti, R. S. K. (1994). New Engl. J . Med. 331, 74-80. O’Neill, B. P., and Illig, J. J. (1989). Mayo Clin. Proc. 64, 1005-1020. Pallesen, G., Hamilton-Dutoit, S. J., Rowe, M., and Young, L. S. (1991). Lancet 337, 320322. Papadopoulos, N. M., Costello, R., and Moutsopoulos, H. M. (1988). Clin. Chem. 34, 973975. Paulus, W., Jellinger, K., Hallas, C., Ott, G., and Muller-Hermelink, H. K. (1993). Neurology 43, 1591-1593.

150

GIANLUCA GAIDANO AND RICCARDO DALLA-FAVERA

Pauza, 'C. D., Galindo, J., and Richman, D. D. (1988).J . C'ir-ol.62, 3358-3564. Pedersen, C., Gerstoft, J., Lundgren, J . D., SkinhGj, P., Bottzauw, J., Geisler, C., HamiltonDutnit, S. J , , Thorsen, S., Lisse, I.. Ralfkiaer, E., and Pallesen, C. (1991).Eur. J. Cancer 27, 1416-1423. Peiper, S. C., Myers, J. L., Broussard. E. E., and Sixbe7,J. W. ( 1990).Arch. Pattivl. Lab. Med. 114, 71 1-714. Pelicci. P . G . , Knowles, D. M., Arlin, Z. A., Wieczorek, R., Luciw, P., Dina, D., Basilico, C., and Dab-Favera, R. (1986a). J . Exp. ,Wed. 164, 2049-2076. Pelicci, P . G . , Knowles, D. M., hlagrath. I . T., and Dalla-Favera, R. (1986b). Prvc. Natl. A c d . Srz. USA 83, 2984-2988. Pelstring, R. J., Zellmer, R. B., Sulak, B. E., Banks, P. M., and Clare. N . (1991).Cancer 67, 1865- 1873. Penn, 1. (1978). , k / r c Coricw ReJ. 28, 31-61. Penn, I. (1981). Clzn. Exp. Inzrnunol. 46, 459-474. Penn, I. (1988). .4nnu. Reii. Med. 39, 63-73. Penn, 1. (1990).AVezisEngl. J . Med. 323, 1767- 1769. Peters, B. S., Beck, E. J., Coleman, D. G ., Wadsworth, M. J. H., MrGuinness, O., Harris, J. R. W.. a n d Pinching, A. J. (1991).Br. Med. J . 302, 203-207. Peterson, R. D., Kelly, W.D., and Good, R. A. (1964). Lancet 1, 1189- 1 193. P1uda.J. hI., Yarchoan, R.,Jaffe, E. S.. Feuerstein, I., Solomon, D., Steinberg, S., Wyvill, K., Raubitschek. A,, Katz, D., and Broder, S. (1990).Ann. Intern. Med. 113, 276-282. Pluda,,l. M., Venzon, D., Tosato, G., Lietzau,J., Wyvill, K., Nelson, D. L., Jaffe, E. S., Karp, J . E.. Brocier. S., and Yarchoan, R. (1993).J. Clirz. Onto[. 11, 1099-1 107. Pollack, K.S . , Sklarin, N . T., Rao, S., and Divon, hI. Y. (1993).Obstet. Gynecol. 81,856-857. Prior, E.. Goldberg, A. F., Conjalka, M. S., Chapman, W. E., Tay, S., and Ames, E. D. (1986).Am. ,/.Rlrd. 81, 1085-1088. Purtilo. D. T.. Strobach, R. S., Okano. M., and Davis, J. R. (1992). Lab. Inve.\t. 67, 5-23. Raab-Traub, N.,and Flynn, K. (1986). Cell 47, 883-889. Rabkin, C. S.. Biggar, R. J., and Horni, J. W. (1991). I n t . J . Cancer 47, 692-696. Ragni, hi. V., Belle. S. H.. Jaffe, R. X., Suerstein, S. L., Bass, D. C., McMillan, C. W., Lovrien, E. M'., Aledort, L. M.. Kisker, C. T., Stabler, S. P., Hoots, W. K., Hilgartner, M. R'.. Cox-Gill, 1.. Buchanan. G. R.. Sanders, N. L., Brettler. D. B., Baron, L. E., Goldsmith,J. C.. Ewenstein, B., Smith, K.J., Green, D., Addiego, J. E., and Kingsley, L. A. (1993). Blood 81, 1889-1897. Ranisay, A. D., Smith, W. J., and Isaacson, P. J. (1988). Am. J . Surg. Pathol. 12, 433-443. Raphael, B. G.. and Knowles, D. XI. (1990). Setnin. Oricol. 17, 361-366. Raphael, 51.. <;entilhomme,O., Tulliez, M., Byron, P.-A,, Diebold, J., and the French Study Group of- Pathology for Hurnan Inimunodeficiency Virus-Associated 'luniors (1991). .4rrh. Pathol. Lab. .%,frd. 115, 15-20, Rechavi. (;., Ben-Basset, M.. Berkowirz, hl.. Martinowitz, U., Brok-Sirnoni, F., Neuniann, Y., Vansover, A., Gottlieb-Stematsky. T., and Ramot, B. (1987). Blood 70, 1713-1717. Kee, H. J . . Stranchen, J. A , , Khan, A. A . . Gold, J. E.. Crowley, J. P.. Kahn, H., and Zalusky, R. (1991). (iurzcei 67, 1614-1621. Reynolds, P., Saunders, 1.. D., Layefsky, 51. E., and Lenip, G. F. (1993). Am. J . Epzdemzol. 137, 19-30. Riboldi, P.. Gaidano, G.. Schettino, E. LV., Steger, T. (;., Knowles, D. M., Dalla-Favera, R., and Casali, 1'. (1994). Bkood 83, 29.52-2961. Rieckniann, P., Poli, G.. Kehrl,J. H.. and Fauci, A. S. (IYUl).J. E x p itled. 173, 1-.5 Roithman. S., Tourani, J . L f . , and Andrieu, J. 31. (1991). Lancet 338, 884-885.

MOLECULAR PATHOGENESIS OF AIDS-RELATED LYMPHOMAS

15 1

Roncella, S., Di Celle, P. F., Cutrona, G., Carbone, A., Sessarego, M., Landonio, G., FOB, R., Rowe, M., and Ferrarini, M. (1993). Leuk. Lymphoma 9, 157-164. Rooney, C. M., Gregory, C. D., Rowe, M., Finerty, S., Edwards, C., Rupani, H., and Rickinson, A. B. (1986).J. Natl. Cancer Inst. 77, 681-687. Ross, R., Dworsky, R., Paganini-Hill, A., Levine, A. M., and Mack, T. (1985). Br. J. Cancer 52,785-787. Rousset, F., Garcia, E., Defrance, T., Pkronne, C., Hsu, D.-H., Kastelein, R., Moore, K. W., and Bancherau, J. (1992). Proc. Natl. Acad. Sci. USA 89, 1890-1893. Rowe, D. T., Rowe, M., Evan, G. I., Wallace, L. E., Farrel, P. J., and Rickinson, A. B. (1986a). EMBO J. 5,2599-2607. Rowe, M., Rooney, C. M., Edwards, C. F., Lenoir, G. M., and Rickinson, A. B. (1986b). Znt. J. Cancer 37, 367-374. Rowe, M., Rowe, D. T., Gregory, C. D., Young, L. S., Farrell, P. J., Rupani, H., and Rickinson, A. B. (1987). EMBO J. 6, 2743-2751. Rowe, M., Young, L. S., Crocker, J., Stokes, H., Henderson, S., and Rickinson, A. B. (199 1). J . Exp. Med. 173, 147-158. Rowe, M., Lear, A. L., Croom-Carter, D., Davies, A. H., and Rickinson, A. B. (1992).J . Virol. 66, 122-131. Safai, B., Diaz, B., and Schwartz, J. (1992). Cancer J. Clin. 42, 74-95. Sa!glio, G., Borrello, M. G., Guerrasio, A., Sozzi, G., Serra, A,, Francia di Celle, P., Foa, R., Ferrarini, M., Roncella, S., Borgna Pignatti, C., Marradi, P., Izzo, P., Soler, J., and Pierotti, M. (1993). Genes Chromosomes Cancer 8, 1-7. Scala, G., Quinto, I., Ruocco, M. R., Arcucci, A., Mallardo, M., Caretto, P., Forni, G., and Venuta, S. (1990).J. Exp. Med. 172, 61-68. Schnittman, S. M., Lane, H. C., Higgins, S. E., Folks, T., and Fauci, A. S., (1986). Science 233, 1084-1086. Schoeppel, J. L., Hoppe, R. T., Dorfman, R. F., Homing, S. J., Collier, A. C., Chew, T. G., and Weiss, L. M. (1985). Ann. Intern. Med. 102, 68-70. Schwab, G., Siegall, C. B., Aarden, L. A., Neckers, L. M., and Nordan, R. P. (1991). Blood 77,587-593. Seremetis, S., Inghirami, G., Ferrero, D., Newcomb, E. W., Knowles, D. K., Dotto, G.-P., and Dalla-Favera, R. (1989). Science 243, 660-663. Serraino, D., Salamina, G., Franceschi, S., Dubois, La Vecchia, C., Brunet, J. B., and Ancelle-Park, R. A. (1992). Br. J. Cancer 66, 912-916. Serrano, M., Bellas, C., Campo, E., Ribera, J., Martin, C., Rubio, R., Ruiz, C., Ocano, I., Buzon, L., Yebra, M., Font, M., and Martinez, A. (1990). Cancer 65, 2248-2254. Shibata, D., Brynes, R. K., Nathwani, B., Kwok, S., Shinsky, J., and Arnheim, N. (1989). Am. J . Pathol. 135, 697-702. Shibata, D., veiss, L. M., Nathwani, B. N., Brynes, R. K., and Levine, A. M. (1991). Blood 77, 1527-1533. Shibata, D., Weiss, L. M., Hernandez, A. M., Nathwani, B. N., Bernstein, L., and Levine, A. M. (1993). Blood 81, 2102-2109. Shiramizu, B., Barriga, F., Neequaye, J., Jafri, A., Dalla-Favera, R., Neri, A., Gutierrez, M., Levine, P., and Magrath, I. T. (1991). Blood 77, 1516-1526. Shiramizu, B., Herndier, B. G., and McGrath, M. S. (1994). Cancer Res. 54, 2069-2072. Silberstein, L. E., Jefferies, L. C., Goldman, J., Friedman, D., Moore, J. S., Nowell, P. C., Roelcke, D., Pruzanski, W., Roudier, J., and Silverman, G. J. (1991). Blood 78, 23722386. So, Y. T., Beckstaed, J. H., and Davis, R. L. (1986). Ann. Neurol. 20, 566-572.

152

GIANLECA GAIDANO AND RICCARDO DALLA-FAMRA

Stein, H., Mason, D. Y.,Gerdes, J., O'Chnor, N., Wainscoat, J., Pallesen, G., Gatter, K., Falini, B., Delsol, G., Lemke, H., Schwarting, R., and Lennert, K. (1985).E M 66,848858. Subar, M.. Neri. A., Inghirami, G., Knowles, D. M., and Dalla-Favera, R. (1988).Blood 72, 667-67 1. Suniaya, C.V., Boswell, R. N., Ench, J., Kisner, D., Hersh. E., Reuben, J., and Mansell, P. (1986).J./njer/. DL..154,864-870. Swinnen, L. J., Costanzo-Nordin, hi. R., Fisher, S. G., OSullivan, E. J., Johnson, M. R., Heroux, A. L., Ditikes, G. J.. Pifarre, R., and Fisher, R. I. (1990).New Ei1gf.J. Med. 323, 1723-1 728. Tanner, J.. and Tosato, G. (1991).J.Clin. I i i t ~ s t .88,239-247. Taub. R., Kirsch, I., and Morton, C. (1982).Pror. Natl. Acad. Sri. USA 79,7837-7841. Tohyama, hi., Karasuyama, H., and Tada, T. (1990).J.Exp. Mid. 171,389-400. Torelli, G., Marasca, R., Luppi, M.,Selleri, L., Ferrari, S., Narni, F., Mariano, M. T., Federico, M.,Ckccherini-Nelli, L., Bendinelli, M.,Montagnani, G., Montorsi, M.,and Artusi, T. (1991).E M 77,2251-2258. Lreki, Y., Goldfarb, I. S., Harindranath. N., Care, M., Koprowski, H., Notkins, A. L., and Casali, P. (1990).J. Exp. Med. 171, 19-34. Unger, P. D.. and Strauchen, J. A. (1986).Canrer 58,821-825. Vieira. P., de Waal-Malefyt, R.. Dang, %i.-N.,Johnson, K. E., Kastelein, R., Fiorentino, D. F., de Vries, J. E., Roncarolo. M-G, Mossmnann, T. R.. and Moore, K. W.(1991). Proc. Satl. h a d . Sri. t!SA 88, 1172-1 176. Volderbing, P. A.. Lagakos, S. W.,Koch, Xi. A., Pettinelli, C., Myers, M. W., Booth, D. K.. Balfour. H. H., Reichman, R. C., Bartlett, J. A., Hirsch, M. S., Murphy, R. L., Hardy, W. D.,Soeiro, R., Fischl, M. A., Bartlett, J. G., hferigan, T. C., Hyslop, N. E.,Richniann, D. D., Valentine. F. T., and Core);, L. C. (1990).New Etrgl. J. Med. 322, 941-949. ron Gunten, C. F., and von Roenn, J. H. (1992).Curr.Vpin. Vncol. 4,894-899. Wang, D., Liebowio.. D., and Kieff, E. (1985).Cill43,831-840. Wang, F.. Gregory, C. D.. Sample, C., b w e . 31.. Liebowiu. D.. Murray, R., Rickinson, A., and Kieff. E. f I!)W). J. Vird 64,2309-23 18. Wang-Peng, J.. Lee, E. C., Sieverts. H., and Magrath, 1. T.(1984).E I d 63,812-822. Wilkes, M.,Fortin, A. H., Felix. J. C., C d w i n , T. A., and Thonipson, W.G. (1988).Luncet 2, 85-88. Yarchoan, R.. Redfield, R. R., and Broder, S. (1986). J. Clin. /tivest. 78,439-447. Yarchoan, R.. Venzon, D. J., Pluda, J. M., Lietzau, J., Wynill, K. M., Tsiatis, A. A., Steinberg, s. hf.. and Broder, s. (1991).A m h t . M i d . 115, 184-189. Ye, B. H.. Rao, P. H.. Chaganti, R. S. K.,and Dalla-Favera, R.(1993a).Cancer Res. 53,27322735. Ye, B. H., Lista, F., Lo Coco, F., Iinc)wles. D. hl.. Chaganti, R. S. K., and Dalla-Favera, R. (1993b). Srieirrr 262, 747-7.50. Young, L., Alfieri, C.. Hennes);, K., Evans, H.. O'Hara. C., Anderson, C., Rip,J., Shapiro, R. S., Rickinson. A., Kieff. E., and Cohen, J. Y. (1989).i\'ew Engl. J. Med. 321, 10001085.

Zelenetz, A. D., Chen, T. T., and Levy, R. (1992).J.Exp. Med. 176, 1137-1 148. Ziegler, J. I,.. Drew, W. L.. Miner, R. C., Rosenbauni, E., Lennette, E. T., Shillitoe, E., Chssavant. C.,Mintz, L., Gershow, J.. Greenspan, J., Beckstead, J.. and Yamamoto, K. ( 1982).Lawet 2, 631-633. Ziegler, J. L., Beckstead, J. A., Volderbing. P. A.. Abrams, D. I., Levine, A. M., Lukes, R. J.. ('311. 1'. S., Burkes. R. L.. Xleyer. P. R.. Metroka. C.E., !tIouradiaci.J., Moore. A., Riggs,

MOLECULAK PATHOGENESIS OF AIDS-KELATED LYMPHOMAS

153

S. A,, Butler, J. J., Cabanillas, F. C., Hersh, E., Newell, G. R., Laubenstein, L. J., Knowles, D., Odajnyk, C., Raphael, B . , Koziner, B., Urmacher, C., and Clarkson, B. D. (1984). New Engl. J. Med. 311, 565-570. zur Hausen, H . , Schulte-Holthausen, H., Klein, G., Henle, W., Henle, G., Clifford, P., and Santesson, L. (1970). Nature 228, 1056-1058.

This Page Intentionally Left Blank

HLA CLASS I ANTIGENS IN HUMAN TUMORS Federico Garrido, Teresa Cabrera, Miguei Angel Lopez-Nevof and Francisco Ruiz-Cabello Servicb de Anelisis Cllnicos e Inrnunologia, Hospital Virgen de las Nieves, Universidad de Grenada, 18014 Granada, Spain

I. Introduction 1 1 . HLA Class I Antigen Expression in Normal Tissues A. MHC Class I Gene Regulation B. Assembly of an HLA Molecule C . Tissue Distribution 111. Role of MHC in T and NK Cell Recognition A. T Cells B. NK Cells IV. HLA and Tumor Antigens V. Alterations of HLA Class I Expression in Human Tumors A. Types of Alterations B. Frequency C. Timing D. Mechanisms of Alteration of MHC Class I Expression in Tumor Cells E. Why Do MHC Losses Occur? VI. Clinical Implications A. HLA-Negative Tumors and Prognosis B. HLA Genes in HLA Active Immunotherapy VII. Conclusions References

I. introduction

T h e major histocompatibility complex (MHC) constitutes a set of genes that synthesize products specializing in the processing and presentation of endogenous and exogenous antigens to the immune system (Germain and Margulies, 1993). These antigens eventually are reduced to small peptides in the endoplasmic reticulum of the cell and transported to the cell membrane together with a particular MHC allele to interact with a T-cell receptor-CD3 complex (Townsend et al., 1986, 1989). Four major categories of genes are involved in this genetically controlled process: class I, class 11, proteasome, and transported genes (Campbell and Trowsdale, 1993). A major characteristic of these genes is their very high degree of polymorphism (Parham et al., 1989). In humans, polyI55 ADVANCES IN CANChK RESEARCH, VOL. 67

I:iipyrighl 0 19!1.5 b y k d r h l i c Press. 111~. ,411 rights of repndurrioti in any form reserved.

156

FEDERICO GARRIDO E T A L .

morphism is concentrated mainly in class I and I1 genes, while in other species proteasomes and transporter genes are also polymorphic (Powis et a/., 1992). Polymorphism provides an enormous diversity of antigenic peptides that can potentially be presented, conferring to a given species the possibility of generating an immune response to a particular antigen, even though a given individual might not have the set of alleles required to do so. However, MHC molecules in mice and humans were not discovered through their physiological functions, but rather through their capacity to produce strong alloimmune reactions that give rise to alloantibodies and alloreactive T cells; these immune effectors have been the tools used to define the molecular organization of MHC antigens and genes (Gorer, 1937; Snell, 1958; Dausset, 1958). Due to the central role of MHC molecules in antigen presentation, including tumor antigens, any alteration in HLA molecule expression may have profound implications for tumor development. The down-regulation of all or some HLA alleles in a particular individual by whatever mechanism breaks the polymorphism and abolishes the capacity to present antigens through MHC products. Cell transformation is associated in many cases with MHC down-regulation, as has been shown in adeno 12 virus infection (Ad12) (Schrier et al., 1983), facilitating the ability of the virus to reproduce in the infected individual, as compared to some nononcogenic viruses that require the appropriate combination of MHC alleles to find the nonresponding individual (Rowtand-Jones~t at., 1995). Other intracellular pathogens such as Plusmodium, which is responsible for malaria infections, also display MHCdriven immunoselection similar to what has been detected in central African populations (Hill rt al., 1992). Virus infections nevertheless are associated in many instances with cytokine production and up-regulation of MHC products, suggesting a host mechanism to enhance immune responses (Rinaldo, 1994; Lehky et d., 1994). The first description of the loss of an H-2Kk private specificity was reported in the Gardener lymphoma, derived from a C3H mouse in the laboratory of Hilliard Festenstein in 1976 (Garrido rt ul., 1976; Garrido and Festenstein, 1976; Festenstein, 1987; Festenstein et al., 1977). It was soon realized that newly induced niouse tumors also lacked H-2 private and public specificities. MCG4, a Balbic tumor, was found not to express the H-2D.4 private and H-2.3, H-2.8, and H-2.13 public specificities iGarrido et ul.. 1979). One particular AKR tumor cell line, designated K36.16, was studied in detail. These cells did not express the K k antigen on the cell surface, were resistant to killing by AKR anti-MuLV cytotoxic lymphocytes 112 d r o , and always developed tumors in immunocompetent ,4KR mice (Schmidt ef al., 1979). The transfection and cell surface expres-

HLA CLASS I ANTIGENS IN HUMAN TUMORS

157

sion of an H-2Kk gene in the K36 (H-2Kknegative) lymphoma inhibited the syngeneic growth of this tumor (Hui et al., 1984) and rendered the mice immune to challenge with the parental line K36 (H-2Kk negative). Further information about the possible role of these H-2 class I alterations in local tumor growth and metastasis accumulated steadily. Other mouse tumor cell lines were transfected with H-2 class I genes and tested for the inhibition of syngeneic growth. Studies with 3-methylcholanthreneinduced T10 sarcoma, of (C3H)Ieb X C57BL/6)F1origin, demonstrated that transfection of the Kk or Kb gene into H-2K-negative parental cells reduced their ability to grow and prevented metastasis (Wallich et al., 1985). The absence of K products, but the normal presence of H-2D (De Baetselier et al., 1980; Katzav et al., 1985), correlated positively with metastatic potential. Similar results were obtained in other experimental tumor models (Tanaka et al., 1985; Gopas et al., 1989; Algarra et al., 1989; Perez et al., 1990). The production and characterization of monoclonal antibodies (mAb’s) against HLA and other cell surface antigens (Barnstable et al., 1978) made it possible to analyze HLA expression in human tumor cell lines. This led to the significant finding that human tumors also displayed altered HLA phenotypes (Pellegrino et al., 1977; Brodsky et al., 1979; Pollack et al., 1980; Festenstein and Schmidt, 1981; Festenstein and Garrrdo, 1986; Garrido, 1991). This chapter will focus on the abnormal MHC expression detected in human tumors, as well as on the biological role that these alterations may have in tumor development. The potential therapeutic implications of these discoveries will also be discussed. II. HLA Class I Antigen Expression in Normal Tissues A . MHC CLASSI GENEREGULATION

MHC class I genes are expressed in most somatic cells, are developmentally regulated, and are inducible by cytokines (Burke et al., 1989; Fellous et al., 1982; Collins et al., 1986). MHC class I antigens are expressed at tissue-specific levels (Harris and Gill, 1986), with the highest levels occurring in lymphoid tissues. 1. Cis-Acting Regulatory Sequences

Transcription of the class I gene is controlled by the conserved cisacting sequences that bind cellular factors (Fig. 1) (Singer and Maguire, 1‘390; David-Watine et al., 1990; Ting and Baldwin, 1993). The first

.

CRE

II

I

' m -203

a

0 167

.161

-6 1

-139

H-2RIIBP CREB AP-1

COUP-TF

H2TFl I . I

_ _ _ - ,.

.

.

..

Enhancer A

ICS

-

J

'.

-

Enhancer B . I

FIG. 1. Map o f the 5' class I region of the niouse MHC. Direction arid initiation of transcription are shown by arrow. T h e posiiions o f the regulatory element enhancer A , interferon consensus element (ICS), and enhancer B are indicated and the transcriptional factors that interact with them.

HLA CLASS I ANTIGENS IN HUMAN TUMORS

159

region is located 20 bp upstream in the vicinity of the transcription start site and forms the TATA box promoter element. The second common region-the CCAAT box-is located upstream of the TATA box. These regions are bound by general transcription factors and are associated with basal levels of transcription of all RNA polymerase II-transcribed genes (Saltzman and Weinman, 1989). The other region comprises elements located upstream of the core promoter and is responsible for the MHC class I gene-specific stimulation of the level of transcription. This region is designated the class I regulatory element (CRE) or enhancer A (Kimura et al., 1986). Analysis of the 5’ flanking regions of class I genes demonstrated the importance of this DNA element in the regulation of class I gene expression, as mutations in these elements interfere with expression (Kimura et al., 1986). The CRE (161-203 bp) is juxtaposed to another regulatory element, the interferon consensus sequence (ICS) (139-167bp) (Israel et al., 1986; Baldwin and Sharp, 1987). In addition, downstream of enhancer A, enhancer B has also been identified; this sequence includes the a-region. This element is active in B cells but not in fibroblasts (Dey et al., 1992). At present, the role of enhancer B in MHC class I regulation is not clear. In addition to these DNA elements, which are predominantly positive sequences, a negative regulatory element located between 679 and 771 bp upstream of the swine MHC genes has been described. This “silencer” element mediates direct repression’ and reduces downstream promoter activity (Weissman and Singer, 1991). 2 . Class I Promoter DNA-BindingFactors

The CRE is critical in determining the level of constitutive class I expression. In the MHC CRE there are at least three sequences (regions 1-111) that bind different nuclear protein factors. Region I contains a symmetrical palindrome and is occupied in tissues that express high levels of MHC class I genes. The region I (RI) element is not occupied with DNA transcriptional factors in HLA class I-negative tissues, whereas these DNA sequence elements interact with many factors in HLApositive tissues (Burke et al., 1989). Several different factors have been identified that bind to the RI region, including KBF1, KBF2, NF-KB, and H2TF1 (Singer and Maguire, 1990). In response to several signals, the transcriptional factor NF-KBbinds to KBenhancer motifs found in a variety of genes (Beuerle, 1991). NF-KB DNA-binding activity involves two subunits, a p50 protein (Ghosh et al., 1990) and a p65 protein (Nolan et al., 1991). Both proteins share a highly conserved N-terminal region and homology with the viral oncogene product v-re1 and its cellular homologue c-re1 (Ghosh et al., 1990). cDNA cloning studies showed that

160

FEDERICO GARRIDO E T AL.

the p50 subunit is equivalent to KBFl (Yano et ul., 1987; Kieran et al., 1990). Although (p50), has been implicated in the maintenance of basal levels of M HC expression on cells, this honiodimer acts as a repressor of H-2Kh gene expression in metastatic tumor cells (Plaksin et ul., 1993). H2TFl is a ubiquitous MHC class I transcription factor that has been identified with N F - K Band ~ also displays high-affinity binding to MHC KBsites (Potter et al., 1993). Region I1 factors are detected in tissues irrespective of MHC class I gene expression (Burke et al., 1989). Binding to the RII element involves positive, H-PRIIBP, and negative transcriptional factors (COUP-TF). H-2RIIBP is a member of the nuclear hormone receptor superfamily that activates MHC class I genes in response to several stimuli (Hamada et ul., 1989; Marks et al., 1992). H-PRIIBP binds to both the estrogen response element (ERE), AGG7 ,.iNNTGACCT, and the MHC class I regulatory element, RII, which shares a sequence similar to the halfsequence of ERE, TGAGGTCA. T h e RII region acts as a negative element in Ad 12-transformed cells, and COUP-TF-binding activity correlates with down-regulation of class I transcription (Liu et al., 1994). Proteins that can interact with K Bsequences interfere in zjitro with binding to the RII elements in human breast cancer cell lines (Rodriguez et ul., 1994).

3 . Regulation b j Interferons Interferons (IFN) alp and y bind to cell surface receptors and induce transcription of overlapping sets of genes such as the MHC class I genes. I F N - a l p and IFN-y act through distinct cell surface receptors (Pestka et ul., 1987) and lead to the induction of multiple ICS-binding factors (Singer and Maguire, 1990). However, the time course of IFN-a-induced transcription suggests that one factor, called interferon-stimulated gene factor-3 (ISGFS), acts as the primary positive regulator (Darnel1 et al., 1994). In fact, transcriptional activation b y IFN-y of genes containing the ICS element is mediated through posttranscriptional activation of the ISGFS complex. ISGF3 is rapidly translocated to the nucleus in response to I F N - d P and binds specifically to ICS. Although IFN-a and IFN-y bind different receptors, p9 1-a subunit of ISGF3-is activated, which induces tyrosine phosphorylation (Shindler et al., 1992; Schuai et al., 1992). T h e tyrosine kinase JAKl is required for the phosphorylation of p91 in response to IFNs-a or -y (Muller et ul., 1993). T h e p91 protein can function in transcriptional activation in two different ways (Schuai et al., 1992). In cells treated with IFN-a, the p91 protein participates in the complex ISRE, whereas in cells treated with IFN-y, this phosphorylated

HLA CLASS I ANTIGENS I N HUMAN TUMORS

161

protein binds to a different DNA element, GAS. p91 acts at the ISGF2GAS interface to mediate IFN-a and IFN-y response. ISGF2/IRF1 is then synthesized in cells treated with IFN-a or IFN-y and acts as a transcriptional activator of MHC class I genes. The IRF family includes ISGF2/IRFl (Miyamoto et al., 1988) and ISGFlIIRF2 (Harada et al., 1989), but whereas ISGF2/IRFl acts as a positive-acting protein and is sufficient to transactive transfected and endogenous class I genes (Chang el al., 1992), IRF2 may be a repressor. EL. ASSEMBLY OF AN HLA MOLECULE

HLA class I molecules are ternary complexes formed by a heavy chain, a light chain of P2-microglobulin (Pam), and a peptide (Barber and Parham, 1993). T h e main function of these complexes is to present antigenic peptide to CD8+ cytotoxic lymphocyte (CTL) (Townsend and ELodmer, 1989). HLA class I molecules are assembled in the lumen of the endoplasmic reticulum (ER), but only if all three components are well complexed, they are then transported through the Golgi apparatus and subsequently expressed on the membrane (Monaco, 1992). 1 . Heavy Chain

Classical HLA class I heavy chains (H-chains) are polymorphic glycosylated proteins of 46 kDa, encoded by three different loci (HLA-A, -B, and -C) comprising more than 100 alleles. After signal sequencedependent insertion in the endoplasmic reticulum membrane, HLA class I mRNA begins to be translated and the protein glycosylated (Ploegh et al., 1979). The HLA class I heavy chain has a short cytoplasmatic tail, a transmembrane domain, and three extracellular domains designated a 1, a2, and a3. It is on the a 1 and a 2 domains that the polymorphic residues are located (Bjorkman and Parham, 1990). When translation is completed, HLA class I H-chain is ligated to a chaperon, calnexin (IPgO), and folded; this avoids the aggregation of unfolded chains (Hochstenbach et al., 1992). The folding process requires the formation of disulfide bonds on a3 and 012, which is facilitated by the selective import of oxidized glutathione into the ER (Hwang et al., 1992). It is also dependent on Ca2+, the concentration of which is normally high in the ER (Sambrook, 1990). 2. &-Microglobulin Pem, a soluble 12-kDa protein, associates with H-chain soon after translation to confer stability and changes in H-chain confarmation

162

FEDERICO GARRIDO E T AL.

(Townsend et al., 1989). A crystallographic model of HLA class I molecules shows a niembrane-proximal region with Prim and the a 3 domain and a menibrane-distal region that includes the a 1 and a2 domains. Pem has a three-dimensional structure similar to that of the a 3 domain and can be included in the immunoglobulin superfamily. Contact between Pam and the a3 domain, in contrast with immunoglobulin constant domains, is asymmetrical. Prim is positioned upward, supporting the a l / a 2 platform. In this position, P2m can interact with all three H-chain domains (B.jorknian et al., 1987). The a 1 and a2 domains form the most critical H-chain structure: the peptide-binding groove. This site is delimited by an a-helix from each domain and has a platform of eight antiparallel P-pleated strands, four from each domain. Polymorphic HLA class I residues are distributed in this cleft on the a-helix and the P-platform (Saper et al., 1991). Although H-chain and &ni binary complexes without peptide can be formed, their turnover is very fast, and their presence on the cytoplasmic membrane is reduced. T h e surface expression of these molecules, called empy HLA class I, can be improved by reducing the culture temperature or loading cells with exogenous peptides (Ljunggren et al., 1990). Therefore, t o achieve normal class I expression, the cleft on a1 and a2 domains must he filled with peptides. 3 . The Protemonie

Peptides are generated from cytosol proteins mainly by a 26s protease (Rechsteiner et al., 1993) after binding to ubiquitin (Michalek et al., 1993). A 20s protease called a proteasome is the proteolytic unit of 26s protease and has a cylindrical structure with four stacked rings (Djaballah et nl., 1993). Each ring has six or seven subunits, named a in the outer rings and P in the inner ones (Kopp et al., 1993). Two proteasome P-subunits, LMP-2 and LMP-7, are encoded by two genes that reside in the HLA class I1 region between the DQ and DP subregions, intimately associated with transporter genes (TAP) (Glynne et al., 1991; Kelly et al., 1991). LMP-2 and LMP-7 are induced by y-interferon (Yang et al., 1992), and in mutant lymphoblastoid cell lines, fibroblasts and normal epidermal cells are replaced by two homologous proteins: delta and MB1. Delta and hlB1 genes map on 17q13 and 14q11.2-14q12 chromosomal bands, respectively (Belich et al., 1994). However, LMP2 and LMP7 allow the proteasome to more efficiently cleave peptides behind basic and hydrophobic residues (Driscoll et al., 1993). This facilitates the production of peptides with carboxyl termini suitable for binding to the HLA class I groove.

HLA CLASS I ANTIGENS IN HUMAN TUMORS

163

4 . Peptde Transporters

From the cytosol, peptides are translocated to the ER lumen by specialized proteins called peptide transporters (TAP). There are two tap isoforms, TAP-1 and TAP-2, encoded by two different genes (Spies et al., 1990).These proteins belong to the ATP-bindingcassette (ABC)transporter family (Trowsdale et al., 1990). They form heterodimers that are predicted to cross the ER membrane six times through the amino-terminal region (Momburg etal., 1994).Loops connecting the membrane-spanning segments are located on the ER and in the cytosol. The transmembrane domain of the heterodimer forms a pore through which the peptides are translocated. The carboxy-terminal domains are located in the cytoplasm and contain the ABC motifs that link ATP hydrolysis with the transport process (Higgins, 1992). Peptide transporters are selective for peptide length and for the carboxy-terminal amino acid. The optimal number of peptide residues ranges from 6 to 15. It is not clear whether peptides are completely cleaved before being translocated by TAP or whether they require additional processing in the ER. The basic or aromatic nature of the carboxy termini determines whether they adapt their side chain to the F pocket on the binding groove (Powis et al., 1992). Each biological species selects for peptides that bind most efficiently to its MHC class I groove. TAP are polymorphic molecules: variability in the rat TAP-2 gene is high, whereas variability in the human TAP-I and TAP-2 genes is low (Colonna et al., 1992).

5. Chaperm It has been hypothesized that peptides from the cytosol bind to several chaperons to avoid degradation. After generation, peptides bind to heat shock protein 70 (hsp 70), which then passes the peptides on to hsp 90; this chaperon, in turn, connects the peptides to the TAP molecule. Inside the ER, peptides bind to hsp 96, which then puts them in contact with the H-chain-P,m complex on the binding groove (Srivastava et al., 1994). Each step is ATP-dependent and is catalyzed through the ATPase activities of hsp 70 and hsp 96 (Li and Srivastava, 1993). hsp 96 complexed with peptides has been used in experiments of cross-priming between allogenic mouse strains (Feldweg and Srivastava, 1993).In vitro assays for peptide transport do not require chaperons, although an important role for these molecules cannot be ruled out when peptide concentration is low (Momburg et al., 1994). An alternative pathway of peptide transport independent of TAP has

164

FEDERICO GARRIDO E T A L .

been proposed in cell lines mutant for TAP. In these cell lines, peptides from signal sequences are bound to HLA class I molecules (Wei and Cresswell, 1992). Signal sequences cross the ER membrane after binding to signal recognition protein (SRP). Inside the ER, these peptides are cleaved by a signal peptidase and can bind to a heavy chain-&m complex. Not all alleles bind with the same affinity to these peptides: in humans, the HLA-A2.1 allele shows the higher affinity, suggesting a selective advantage (Henderson et al., 1992). In other cell lines, TAP-1 can function like a homodimer to transport specific viral peptides (Gabathuler et al., 1994). 6 . Peptides

T h e rules that govern peptide binding to the class I groove have been elucidated by analyses of naturally processed class I-associated peptides (Falk ct al., 1991). Pool sequencing by Edman degradation showed that naturally processed peptides eluted from HLA class I molecules are 8-9 residues long with two anchor motifs at position 2 (1’2) and P8 or P9. ‘These anchor motifs are constant in most peptides that bind to a specific class I allele (Rotzschke et al., 1992). Different technical approaches, such as peptide isolation by HPLC prior to Edman sequencing (Jardetzky et al., 199 1) or microcapillary HPLC in combination with electrospray ionization/tandem mass spectrometry (Hunt et al., 1992), have revealed the features of individual naturally processed peptides. With these methods, 20-40 of all individual peptides are longer than nine residues, and comparative sequence analyses confirmed the presence of anchor motifs (Guo et ul., 1992). If polymorphic residues of class I heavy chain are located on the peptide-binding groove, each allele must have a groove with a specific conformation that selects for a specific set of peptides. Crystallographic analyses of the groove show six pockets, named A-F, where interaction with peptide residues takes place (Garrett et al., 1989). Pockets A and F, which determine peptide orientation, are located at opposite ends of the groove. Pocket A interacts with amino-terminal peptide residues, and pocket F interacts with carboxy-terminal residues through hydrogen bonds from conserved class I heavy chain residues (Madden et al., 1992). ‘The anchor residues on P2 and P8 or P9 are determined by pocket F and pocket B, respectively, which present specific allelic conformations for peptide residue side chains (Carreno et nl., 1993). Polymorphic class I heavy chain residues determine pocket conformation. Other interactions are more relaxed, for example, P3 with pocket D, PG-pocket C, and P7-pocket E. In the peptide residues occupying these pockets, the side chains face into the groove and are not detected by T-cell receptors

HLA CLASS I ANTIGENS I N H U M A N TUMORS

165

(e.g., P2, P3, P6, or P9). The side chains of P1, P4, and P5 peptide residues are directed out of the groove and can be recognized by T-cell receptors (Matsumura et al., 1992). Increasing numbers of naturally processed peptides have been sequenced after their elution from several locus A and B alleles (Engelhard, 1994a). There is no peptide sequence available from HLA-C alleles. Many peptides are of unknown origin. Peptides from ribosomal proteins, related cell-cycle proteins, and signal sequences are relatively frequent. In one experiment, differences between peptides from HLA-2.1 in several lymphoblastoid B and melanoma cell lines were found in only 10% of the total peptide pool, a finding in agreement with the fact that only a small percentage of gene expression is tissue-specific (Engelhard, 1994b). Quantitative analyses predict that one cell can have around 10,000 different peptides in HLA class I complexes. An individual peptide may be represented by 200-400 HLA class I complexes (Huczko et al., 1993). A database of MHC-binding peptides (MHV4.CPEP) contains 4000 peptide sequences, including naturally processed and exogenously defined peptides (Brusic et al., 1994). C. TISSUE DISTRIBUTION

Significant changes in HLA expression in tumor cells are assessed by determining the proportion of neoplastic cells that exhibit differences in inimunohistochemical labeling in comparison with normal tissues from thie same specimen. Most glandular or squamous epithelia, as well as the surrounding connective tissue, express HLA class I antigens (Redman et al., 1984; Natali et al., 1984; Daar et al., 1984). However, the intensity of expression varies between different localizations. For instance, weak expression is detected in endocrine cells of the thyroid, parathyroid, islets of Langerhans, myocardium, skeletal smooth muscle, gastric mucosa, and hepatocytes (Daar et al., 1984; Fernandez et al., 1991; Ferron et al., 1989). Tissues that show very weak or negative expression are the corneal epithelium, duodenal Brunner glands, trophoblastic villi, neurons of the central nervous system, exocrine portion of the pancreas, and acinar cells of the parathyroid gland (Daar et ul., 1984). These data have been obtained by using predominantly immunohistological techniques that are known to be relatively insensitive. It is thus possible that new information will come to light when other molecular biology techniques are used. However, findings suggest that there is a correspondence between the HLA expression detected in solid tumors with immunohistological techniques and the expression observed with F.ACS analysis and reverse trdnscriptase-PCR in cell lines derived from

166

FEDERICO GARRIDO E T A L .

these tumors (Ruiz-Cabello et al., 1995). Nonetheless, it should be emphasized that similar o r identical techniques need to be used to draw meaningful comparisons between HLA expression in a tumor and in its corresponding normal tissue. Classical HLA class I molecules (A, B, C) are not expressed on human spermatozoa (Hass and Nahhas, 1986) or unfertilized or fertilized oocytes (Roberts et al., 1992), nor are they detected on cytotrophoblast or syncytotrophoblast on human placenta (Hunt and Hsi. 1990). This may represent a biological strategy to avoid allogenic immune response from mother to fetus. In contrast, nonclassical HLA class I genes are expressed on the placenta at the cell surface (HLA-G) or at the mRNA level (HLA-E) (Wei and Orr, 1990). Their function is still unknown, although both can bind peptides.

I l l . Role of MHC in T and NK Cell Recognition A. T CELLS In contrast with B lymphocytes, T cells d o not react with isolated antigens. T-cell-specific stimulation requires antigen-presenting cells and antigen processing (Rosenthal and Shevach, 1973). Morever, primed T lymphocytes must be MHC-identical to antigen-presenting cells to be stimulated. This requirement is known as restriction of T-cell response by MHC (Zinkernagel and Doherty, 1974). There are two T-cell subpopulations: CD4+ helper cells restricted by MHC class I1 molecules and CD8+ cytotoxic or suppressor cells restricted by MHC class I molecules (Swain, 1983).T-cell receptor (TCR), a transmembrane heterodimer, is usually divided into two classes, a and (3 and y and 6 (Kronenberg et al., 1986), and both classes are associated to the CD3 complex (Clevers et al., 1988). Studies with a/@CD8+ CTL specific for influenza virus nucleoprotein demonstrated that TCR recognizes short peptides associated to MHC class I molecules. Uninfected target cells expressing the appropriate class I allele can be lysed by specific CTL after incubation with immunogenic peptide. This model, called peptide-scan, has been widely used to define peptide epitopes by exogenous incubation with a wide set of antigen fragments (Townsend and Bodmer, 1989). T h e CTL response against a viral nucleoprotein peptide was the basis for postulating that peptide and MHC class I molecule (heavy chain + Prim) form an intracellular complex that is transported to the membrane and expressed. Crystallographic studies, molecular dissection of the exogenous antigen-processing component (proteasome, TAP), and analy-

HLA CLASS I ANTIGENS I N HUMAN TUMORS

167

ses of naturally processed peptides have provided molecular evidence in support of this hypothesis. Peptide recognition by CTL is a good strateg y for the immune system against pathogens that are not accessible to antibodies when the former are inside the cell. This system permits the elimination of viral infected cells before all viral proteins can be transla ted, preventing viral dissemination (Germain, 1993). Normal cells express class I molecules that present a very wide set of endogenous peptides. CTL must discriminate between self- or non-selfpeptides in order to avoid autoaggression: this is a central key of the immune system. T h e predominant theory explains that T cells are educated in the thymus, where endogenous self-peptides are displayed by autologous MHC molecules on epithelial or other stromal cortical cells. T cells that do not bind any self-peptide MHC molecules die, and those that react are positively selected (von Boehmer, 1994).From this population, autoreactive cells with high affinity are deleted from the medulla or the interface between the cortex and the medulla (negative selection), resulting in a population of effector cells that have intermediate affinity for self-peptide MHC molecules (Nossal, 1994). Some autoreactive clones are energized, and other T cells are unaffected by antigenic encounter (clonal ignorance). A controversial issue is the relative importance of the peptide alone versus the class I heavy chain in thymus selection. The same question arises logically with regard to TCR-foreign epitope interactions. T h e effectiveness of a peptide from a foreign antigen as a CTL epitope is influenced by several factors related to the endogenous antigen plrOCeSSing and presentation pathway. The most critical factors are binding affinity to class I molecules, correct expression of the complex on the cytoplasmatic membrane, and recognition of accessible peptide residues or peptide conformations by TCR (Parham and Barber, 1994). Binding affinity to class I molecules is related to interactions between peptide anchor residues and groove pockets on class I heavy chain, but in ather epitopes, the overall peptide structure is required to maintain binding (Chen et al., 1994). By using algorithms, it is possible to identify-on known protein sequences-peptide candidates to act as CTL epitopes that present anchor residues for a given class I allele (Rotzschke et al., 199 1). Several groups have defined class I-restricted T-cell epitopes (Hill et al., 1992; Pamer et al., 1992), but in other cases this approach was unsuccessful because internal residues abrogated binding or constituted rnore complex and unusual secondary motifs (Elliott et al., 1992; Maryanski et al., 1991). In these cases, quantitative affinity measurements can be performed with biosensor-based technology and surface plasmon resonance (SPR) (Khilko et al., 1993). HPLC isolation and sequencing of peptides eluted from MHC class I

168

FEDERICO GARRIDO ET A L .

molecules is a direct method to define CTL epitopes after incubation with adequate antigen-presenting cells. This procedure is used when the antigenic protein is unknown, as is frequent in antitumoral CTL responses. Peptide sequences can identify the protein that contains the sequence or can be used to make an oligonucleotide for cloning a new gene (Udaka pi al., 1992; Wallny et al., 1992; Slingluff et al., 1993). TCR a i p from CD8+ CTL recognizes accessible side chains from peptide residues protruding out of the groove ( P l , P4, and P5) by its complementary determining region 3 (CDRS) (Sorger et al., 1990). There is no consensus as to whether TCR a/@interacts with residues on a-helices of the binding groove (Davis and Bjorkman, 1988) or only with bound peptide (Claverie and Kourilsky, 1986). T h e iack of crystal structures for TCR-binding MHC class I complexes means that this question must still await a satisfactory answer. Although TCR a/@has low affinity for MHC class I ternary complex, K , = 10-4-10-5 (Weber et al., 1992), low numbers of peptide-MHC complexes are able to trigger a T-cell response (Christnick et al., 1991).In order to reconcile these opposite factors, it has been proposed that the TCR concentration on the APC-T-cell interface is high, which compensates for low affinity, and that T-cell triggering can be achieved by conformational TCR changes instead of TCR cross-linking (Karjalainen, 1994). 'TCR binding to peptide-MHC class I complex is reinforced by CD8 interaction with a3 domains of heavy chain (Salter et al., 1990).Adhesion molecules are also important for CTL-target interaction. Lymphocyte activation (Collins et al., 1994) requires a costimulation signal provided by CD28-B7.1 (Azuma et al., 1993) and CTLA-4-B7.2 interactions (Freeman et al., 1993).CD8+ CTL are under the control of CD4 type I helper cells, which produce IL-2 and y-interferon (Glasebrook and Fitch, 1979). Recently, a possible explanation for the impaired T-cell responses found in animal model and cancer patients has been postulated, evidenced from alterations of signal transduction molecules in T lymphocytes from tumor-bearing mice. Tumor-bearing mice expressed T-cell antigen receptors that contained low amounts of CD3y and completely lacked CD35, which was replaced by the Fc, y-chain (Mizoguchi et al., 1992).

H . K K CELLS NK cells are cytotoxic cells that lack rearranged receptor-like TCR. In the absence of prior stimulation, they can induce the lysis of virus infected and tumor cells. Morphologically, these cells are large granular lymphocytes (LGL) (Trinchieri, 1989). I n uztro maturation assays demonstrated that fetal or neonatal thymus and fetal liver contain precursor

HLA CLASS I ANTIGENS IN HUMAN TUMORS

169

populations that can differentiate to NK to T lineage, depending on cultiire conditions (Rodewald et al., 1992; Poggi et al., 1993). It is not known whether T and N K cells have the same or different cell precursors. N K cells lyse target cells in an unrestricted MHC class I fashion, but NK susceptibility is influenced by the level'of MHC class I. Target cells that do not express MHC class I molecules are the most sensitive to NK lysis (Ljunggren and Karre, 1990). T o explain this relation, it was proposed that class I molecules can induce a negative signal to NK cells or, alternatively, they could mask a target molecule. Subsequent work supported the first hypothesis. The determinant involved in NK recognition of MHC class I molecules resides on the peptide-binding class I a l / a 2 domains (Storkus et al., 1991). HLA class I specificities recognized by NK cells are public, shared by several HLA class I alleles. Four discrete NK receptors have been described that define HLA class I specificities. These receptors are not related with TCR and can be coexpressed in the same NK cell. They induce a negative signal on NK cells that specifically protects targets expressing a given HLA specificity from lysis. The NK resistance induced by HLA molecules is dominant and unidirectional. NK receptors that mediate NK resistance have been characterized by using allogenic NK clones against PHA blast or HLA class I transfectants and with monoclonal antibodies recognizing NK receptors that reverse NK resistance (Moretta et al., 1994a). The first two NK receptors that recognize HLA class I specificities and induce NK resistance belong to the p58 family and define an HLA-C dimorphism (NKl and NK2 specificities) on residues 77 and 80 of HLA class I heavy chain (Colonna et al., 1993). The third N K receptor is Kp43 (CD94), which is specific for several HLA-BW6 alleles (HLA-B7, -B8, and -B14) (Moretta et al., 1994b). The fourth is NKBl, a 70-kDa glycoprotein that recognizes several HLA-Bw4 alleles (HLA-B*510l , -B*580 l , and -B*2705) (Litwin et al., 1994). Another HLA-Bw4-related specificity, named NK3, is located, like NK1 and NK2 HLA-C specificities, on residue 80 or the heavy chain and is related to NKBl specificity (Cella et al., 1994). No HLA-A-associated specificity is recognized by NK cells. Genes from these NK receptors have not been cloned, and there are few structural protein data. p58 receptors are hetero- or homodimers associated with CD4 {-chain (Moretta et al., 1994a). Kp43, also a homodimer, is related to lectin-like molecules (Aramburu et al., 1991), and NKB1 is a glycoprotein. It has been proposed that the physiological role of the N K receptor fior HLA class I is to detect alterations in the conformation or expression of MHC class I molecules (e.g., total, locus or allelic HLA class I losses) induced by intracellular pathogens, such as virus, or by tumoral trans-

170

FEDERICO GARRIDO E 7 A L .

formation. Cells with these alterations eventually lose h'K resistance and can be ivsed (Trinchieri, 1994). IV. HLA and Tumor Antigens

It has been possible to define several genes that code for peptides specifically recognized by cytolytic T cells, which lyse autologous tumor cells. These genes were cloned and sequenced in the murine mastocytoma 1'815 (H-2") and in a human melanoma cell line (MZPMEL). Two major mechanisms were defined that generate peptides recognized by autologous T cells: ( 1 ) a point mutation in the gene, which produces a mutated peptide (Boon and Kellerman, 1977), and (2) a normal (nonmutated) peptide derived from a protein not exposed to the immune system during embryonic development (Boon, 1983). This pioneering work has provided evidence that peptides presented together with the corresponding HLA class I molecule can behave like tumor antigens. It must be kept in mind that tumor rejection antigens have been defined in I J Z Z ~ Oand may or may not correspond to the peptides recognized by cytotoxic T lymphocytes. However, in the P815 murine mastocytoma, the tumor minus subline obtained by in uztro mutagenesis was able to produce a mutated peptide recognized by syngeneic T cells. In this particular tumor model, the tumor rejection antigen, the tumorassociated transplantation antigen (TATA), and the mutated peptide recognized by syngeneic T cells may be the same structure (Boon, 1983). These observations are concordant with findings in tumors induced by oncogenic viruses, in which it has been possible to preimmunize animals against syngeneic tumor growth using a particular combination of peptides derived from the viral genome (Melief, 1991). Once again, these tumor cells were recognized by autologous T cells against these same predefined peptides (Melief and Kast, 1991). These independent experiments suggest that, in some instances, the MHC molecules may be presenting t o T-cell tumor antigens that behave like tumor rejection antigens. If this were the case, it would constitute evidence of the possible role of MHC class I molecules in antitumor response. The high degree of polymorphism displayed by MHC molecules parallels the polymorphism of TATA, which was described in the 1950s in methylcholanthrene-induced experimental tumors (Klein et nl., 1960). However, whether mutated or nonmutated peptides can replace the effects of the T A T A remains to be seen. Net ertheless, other immune effector mechanisms are involved in antitumor immune responses (Garrido and Klein, 1991), and it would be oversimplifying matters to identify TATA with peptides presented by

HLA CLASS I ANTIGENS I N HUMAN TUMORS

171

MHC molecules. There is also evidence that a single tumor cell can present several antigens restricted by different MHC molecules to different clones of cytotoxic T cells. Some of these antigens are immunodominant and can modulate the T-cell response against other minor antigens’ (Seung et al., 1993). In humans, CD8+ CTL clones have been used to define several tumoral antigens on melanoma restricted by HLA class I molecules. MAGE-1 was the first human melanoma antigen described (van der Bruggen et al., 1991): it belongs to a family of at least 12 genes located on chromosome X. Two peptides generated by MAGE- 1 are presented by HLA-A1 (Traversari et al., 1992) and HLA-Cwl601 (van der Bruggen et al., 1994). Another member of the family, MAGE-3, encodes for a peptide also presented by HLA-A1. These genes are also expressed in others tumors and in normal testis. Their physiological function is unknown. In the search for tumor antigens presented by HLA-A2 molecules, three differentiation melanocyte genes that induce an anti-melanoma response in vitro have been described: tyrosinase, gp100/pMel-17, and MART-l/Melan-A (Slingluff et al., 1994). They are also expressed on normal melanocytes. The peptides encoded by these genes are not mutated, a situation similar to MAGE peptides. This raises the question of whether these differentiation antigens are tolerated by the immune system, probably by clonal ignorance, because they are not expressed on the thymus. There is also evidence that the products of activated protooncogenes can be processed by tumor cells and presented to T cells. Mutated peptides derived from mutated ras products can induce specific CTL in some human tumors (Gedde-Dahl et al., 1993). pEi3 and HER-2lneu have also been shown to present peptides that induce CTL responses in vitro (Slingluff et al., 1994). This example indicates the possibility that the different protooncogenes discovered so far may be another important source of peptides to be presented to T cells as tumor antigens. V. Alterations of HLA Class I Expression in Human Tumors

A,. TYPES OF ALTERATIONS With the first description of a mAb able to recognize a conformational epitope of an HLA antigen and able to work in tissue sections, several laboratories started to detect the absence of reactivity in some tumor samples (Pellegrino et al., 1977; Pollack et al., 1980). It soon became evident that a proportion of tumors derived from HLA-positive epithelia did not react with anti-W6/32 o r anti-P,-microglobulin mAb’s (Csiba

172

FEDERICO GARRIDO E T AL.

HLA CLASS I

+++

TOTAL LOSS

A1

cw4

B LO C U S L O SS

A LOCUS LOSS

A1

'/ cw4

cw4

B ALLELIC LOSS

A ALLELIC LOSS

A1

A3

cw2

// cw4

cw4

FIG. 2. Hypothetical alterations in the HLA phenotype of tunior cells derived from a randomly chosen normal cell. Total. locus-specific, and allelic losses are shown.

ef al., 1984; Momburg et al., 1986; Lopez Nevot et al., 1986; Ruiz-Cabello et al., 1987; Ruiter et al., 1991). The absence of cell surface expression was confirmed by different authors when iodine-labeled tumor cell lines were irnmunoprecipitated using the same anti-HLA antibodies that recognized monomorphic determinants. I t was later realized that many different rnechanisms could lead to the interruption of the machinery

173

HLA CLASS I ANTIGENS IN HUMAN TUMORS

required to synthesize a functional HLA molecule. The final consequences of these alterations was the absence of all or some HLA alleles from tumor cells. T h e existence of anti-HLA mAb’s capable of specifically detecting HILA-A and HLA-B locus products (Yang et al., 1984; Burrone et al., 1985; Lozano et al., 1989) helped to define HLA-A and -B locus losses in tuimor cells (Ferron et al., 1989; Lopez-Nevot et al., 1989; Redondo et al., 1991b; Ruiz-Cabello et al., 1991a; Smith, 1991). Different anti-HLA mAb’s that define specific anti-HLA alleles have become available for tissue section studies (Keating et al., 1995). This rapidly increasing list of m.4b’s is helping to define specific HLA allelic losses in tumor tissue sections. Figure 2 summarizes the three major HLA alterations observed in tumor cells, starting from a hypothetical HLA class I-positive cell expressing an HLA-A 1 ,A3/B8,B 18/Cw2,Cw4 phenotype. Total losses, HLA-A and -B locus losses, and HLA-A and -B allelic losses are represeinted in this figure. Frequency

Yo

loo 80

60

40

1

/

I /

1

:I

2o 0 1

Allelic

HLA A,B

Monornorphic mAbs

I

1970

I

t

I

1980

I

I

1990

I

1994

I

2000

H-2

FIG. 3. Graph of the detection of alterations in the HLA phenotype in tumor cells. T h e arrow indicates the first report of the loss of an MHC antigen in a mouse tumor. The po8intsindicate dates when anti-HLA mAbs against monomorphic (1979), A and B locusspecific (1984), and allele-specific determinants (1994) came into use. Foreseeably, u p to 95% of all human tumors may be shown to have some type of alteration in the HLA phlenotype.

174

FEDERICO GARRIDO ET AL.

B. FREQUENCY

With the growing use of mAb's against monomorphic determinants in the late 1970s (Barnstable et al., 1978), HLA alterations began to be detected in human tumors. Many samples derived from different tissues were analyzed, and concordant results were obtained by different groups (Ruiz-Cabello et al., 1987; Ruiter et al., 1991; Garrido and Klein, 1991; Moller et al., 1991). At that time it was thought that only a small proportion of tumor cells (10-20%) contained such HLA alterations (Fig. 3). i n the mid-l980s, several groups started to use anti-HLA-A and -B locus-specific mAbs, and the figure approached 40% in certain tumors. Data from the first analyses with anti-HLA allelic mAb's have become available, and around 60-70% of tumors have been shown to present a particular HLA alteration (Keating et al., 1995). Nevertheless, it is important to emphasize that an extensive panel of anti-HLA allelic mAbs that work in human tumor tissue sections and that cover all (or most) HLA alleles is not yet available. For this reason, it is possible to predict that the percentage of tumors presenting HLA alterations could actually be close to 90-95% (Fig. 3).

c. T I M I N G We have found that HLA alterations detected in different tumor types derived from four major HLA class I-positive tissues (colon, breast, cervix, and larynx) take place when the tumor becomes invasive and starts to metastasize (Fig. 4) (Garrido et al., 1993). Indeed, benign lesions from these epithelia, i.e., colon adenomas (Gutierrez et al., 1990), high-risk proliferative lesions of the breast, cervical intraepithelial neoplasia of the cervix, or squamous papilloma of the larynx, are HLA-positive (Petersen et al., 1993; Garrido et al., 1993; Torres et al., 1993). Although many oncogenes or tumor suppressor genes are activated or inactivated during

Benign tumors

,

In situ carcinomas

1

HLA Class I

,

,nvasion

FIG. 4. 'rime course of-HLA alterations during the multistep process of'tumorigenesis. T&I~. locus-specitic. or allele-specific losses seem t o occur when the tumor becomes invas h e and begins to nietastatize.

175

HLA CLASS I ANTIGENS IN HUMAN TUMORS

these early events (Fearon and Volgestein, 1990), HLA antigens apparently are expressed in a normal way. Even in situ carcinomas, which are histopathologically malignant, are HLA-positive in the tumor populations that we analyzed. HLA alterations seem to occur when the tumor disrupts the basal membrane and becomes invasive. It is also possible th,at HLA-negative variants preexist in premalignant lesions, but cannot be detected until late stages due to the limited sensitivity of the technique used (Garrido et al., 1993).A greater number of premalignant lesions need to be carefully studied in order to substantiate these findings. D. MECHANISMS OF ALTERATION OF MHC CLASSI EXPRESSION IN TUMOR CELLS

Tumor cells escape immune recognition through multiple heterogeneous mechanisms (Table I). Alteration of the expression of the molecules involved in recognition may affect the processing and presentation of TAAs, and abnormalities in HLA class I expression are thought to be

TABLE 1 GENERAL MECHANISMSOF MHC CLASSI DOWN-REGULATION Level of alteration ~

~~

~~

Heavy chain

Mechanism

Referencesa

~

Reduction of MHC class I mRNA levels Methylation Decreased binding of positive trans-acting factors Enhanced binding of negative trans-acting factors Allele or locus-specific down-regulation Enhancer suppression of c-my Genomic loss Interference with assambly and/or transport Retention of MHC class I molecules and peptides Inhibition of glycosylation Enhanced of degradation of heavy chain

1-3 4-7 8-10

11 12-13 14, 15 16

17

P,-Microglobulin

Mutations in the

Prim gene

18-20

TAP

Reduction of TAP-1 and TAP-2 mRNA levels Mutation in TAP genes

21-23 24-28

0 References: 1, Alberti and Herzenberg (1988); 2, Maschek et al. (1989); 3, Bouraut etal. (1993);4, Henseling el al. (1990); 5, Blanchet et al. (1991); 6, Blanchet et al. (1992); 7, Van't Verr et al. (1993); 8, Liu el al. (1 994); 9, Ge el a[. (1994); 10, Howcroft el al. (1 993); 1 1, Versteeg et al. ( 1988); 12, Marincola et al. (1994); 13, Ruiz-Cabelloetal. (1995); 14, Andersson etal. (1985); 15, Korner and Burgert (1994); 16, HIIIIet al. (1994); 17, Beersma et al. (1993); 18, Rosa et al. (1983); 19, DUrso etal. (1991); 20, Bicknell et al. (1994); 21, Restifo et al. (1993); 22, Cromrne et al. (1994); 23, Stern el al. (1994); 24, Townsend et al. (1989); 25, DeMars et al. (1985); 26, Hosken et al. (1990); 27, Cerundolo et at. (1990); 28, Spies et at. (1'391).

176

FEDERICO GARRIDO E T AL.

among the most important of these mechanisms. Any step in the synthesis of class I molecules can be the target for an alteration that results in the absence of a particular molecule. 1. Abnormalitks That Affect M H C Class I Heauy Chain

Numerous mechanisms have been reported that induce complete or selective loss of MHC expression (Moiler and Hammerling, 1992; Garrido Pt d., 1993). Because o f the relatively high frequency of total loss of HLA antigens in tumors, it seems unlikely that mutations or deletions exist in the coding regions of the genes for the heavy chain, as these changes would necessarily lead to multiple mutations. In fact, in cases where total loss of HLA expression is evident, genetic rearrangements of HIA-ABC genes are not found (Lopez Nevot et al., 1987, 1989; Esteban et ul., 1989). However, point mutations, genetic rearrangements, and genomic loss of encoding HLA class I heavy chains genes have been shown to be involved in the loss of expression of one or more alleles. The failure in cell surface expression of H-2K’) in variants of a tumor cell line was caused by a single base change (G to A transition) in exon 3 . T h e mutations involve the substitution of Tyr for Cys, thereby disrupting an intrachain disulfide linkage and abrogating H-2Kt>expression (Zeff et al., 1990). Loss of heavy chain genes has been observed after y-irradiation in B lymphoblastoid cell lines (Kavathas ef al., 1980). Genomic losses have been suggested to explain the high incidence of loss of single HLA-A locus specificities in colon carcinoma cell lines (Browning et al., 1993). L,oss of heterozygosity for HLA alleles was reported in melanoma cell lines (Marincola et ~ l . ,1994). However, i n these studies the original tumor tissue was n o t available for analysis, raising the possibility that such changes may have been due to selection during prolonged in uitro culture. Loss of an HLA haplotype in pancreas cancer tissue and its corresponding tumor-derii.ed cell line has been reported (Kuiz-Cabello P t d., 1995). ‘I’he absence of A30 arid B14 allelic expression from the IMIM-PC;-:! tumor pancreatic cell line is due to a genoniic loss in chromosome 6. .A characteristic of this type of lesion is the absence of response to IFN-y. However, in many cases, reexpression of MHC class I antigens is possible after modulation with IFNs 01’ other cytokines. When -his occurs, regulatory mechanisms common to heavy chain genes and &n. are affected most frequently. I n this context, various studies have confirmed a direct association between cell surface expression and mRNA levels for HLA class I, suggesting transcriptional regulation o f HLA expression (Doyle at a/., 1985; Ruiz-Cabello et al., 1991b; Blanchet Pt ( I / . , 190 1 ).

HLA CLASS I ANTIGENS IN HUMAN TUMORS

177

A cis-acting regulatory mechanism that alters methylation and chromatin structure may affect HLA expression (Alberti and Herzenberg, 1988; Maschek et al., 1989; Bonal et al., 1986). It has been suggested that the MHC class I chromosomal region in the trophoblast tumor cell line JAR can be divided into three domains according to their states of methylation, chromatin conformation, and transcriptional activity. An inverse correlation between state of methylation and transcriptional activity was found for the classical class I polymorphic genes, as well as for HLA-F and HLA-G genes (Boucraut et al., 1993). The constitutive and inducible expression of HLA class I genes is regulated by cis-acting regulatory sequences (David-Watine et al., 1990; Halloran and Madrenas, 1990; Singer and Maguire, 1990). These sequences have been shown to bind trans-acting protein factors associated wilh MHC class I expression (Fig. 1) and related to HLA alterations. In fact, altered binding of the NF-KBIKBF~ regulatory factor to a class I enhancer sequence was described in cell lines lacking class I antigen expression (Henseling et al., 1990; Blanchet et al., 1991, 1992). However, there is an independent mode of transcriptional activation by the p50 (KBF1) and p65 subunits of NF-KB(Fujita et al., 1992).This may explain whiy an excess of KBFl over NF-KBis observed in low-expressing metastatic tumor cells, whereas an excess of NF-KBover KBFl is observed in high-expressing cells (Plaksin et al., 1993). Transcriptional locus-specific factors have also been implicated in tumor cell lines with defective locus expression (Soong and Hui, 1991). In fact, differential regulation of HLA-A versus HLA-B has been described (Burrone et al., 1985; Hakem et d.,1991). The enhancer of HLA-A contains two KB-binding motifs, whereas the HLA-B enhancer has only one (Girdlestone et al., 1993). Altered binding of transcriptional factors may be mediated by onco:genic products. Expression of MHC class I molecules is depressed by both N-myc and c-myc in neuroblastoma and melanoma cell lines (Bernards ct al., 1986; Versteeg et al., 1988). N-myc may act by downregulating the p50 subunit of NF-KB (Van? Veer et al., 1993). N-myc inhibits KBF 1 activity, reducing p50 gene expression and inhibiting NI?-KBactivity. Furthermore, transfection of a p50 expression vector in neuroblastoma cells that express high levels of N-myc leads to restoration of the factor binding to the MHC class I gene enhancer and enhancer activity and reexpression of MHC class I antigens at the cell surface. N-myc alters the expression of protein kinase C (PKC) isoforms (Bernards, 1991); as a result, NF-KB cannot be activated by PKC in N-myctransfected cells. A different mechanism involves c-myc-mediated MHC class I downmodulation. In fact, there is a clear difference between c-myc and N-myc:

178

FEDERICO GARRIDO E T A L .

c-niy expression is related to locus-specific down-modulation, mainly of HLA-B alleles, whereas X-myc switches off all HLA class I molecules. Furthermore, whereas the effect on class I expression by N - r a y is regulated through enhancer A, the effect mediated by c-myc appears to involve different regulatory elements located in the core promoter (Schrier et al., 1991; Schrier and Peltenburg 1993). Aside from the inverse relationship between HLA class I mRNA and c-myr mRNA described in melanoma cell lines, this mechanism is not present in other melanoma, neuroblastoma (Feltner et al., 1989), or non-small-lung carcinoma cell lines (Redondo et al., 199la). This absence suggests the participation of tissue-specific factors in the down-modulation of MHC class I expression by c - m y . MHC class I correlated positively with c$os expression in murine tumor cells (Kushtai et al., 1988; Gaforio et al., 1992).T h e transfection of c-fos converted metastatic clones with low expression of MHC class I antigens into nonmetastatic phenotypes with high expression of these antigens. T h e relation of c-fi~st o MHC class I expression may be explained by the presence of AP-l-binding sites within the regulatory region of MHC class I genes. However, in other cases c-fos did not correlate with MHC class I expression (Barzilay et al., 1987; Ruiz-Cabello et al., 199la). Other data indicate that c-junlAP-1 acts as a negative transacting factor that down-regulates MHC class I gene expression. MHC class I gene expression is negatively regulated by the protooncogene c-jutz Howcroft et ul., 1993), and the overexpression of c-jun results in decreased MHC class I promoter activity and reduces the steady-state level of MHC class I RNA in murine cells. No direct evidence exists for a relationship between alterations in MHC expression and activation of the K-rus oncogene by point mutation (Oliva et al., 1990). Likewise, other studies have failed to observe an). correlation between MHC class I expression and N-rcls o r H-ras activation (Grand et al., 1987; Elliot et al., 1989). However, it has been reported that mutant v-ras induces downregulation of MHC class I expression in murine-transformed fibroblasts (Lu et al., 1991). Inactivation of tumor suppressor genes is believed to be important in the development of malignant transformation. At present, changes in MHC expression have not been related to the loss offunction of tumor suppressor genes. 2. Lack qf HLA Class I Antigen Expression by Defects In p,-iclicroglobulin

&m and peptides are essential for the conformation and transport of the HLA class I molecules to the cell surface (Neefjes et al., 1993). Lack

HLA CLASS I ANTIGENS IN HUMAN TUMORS

179

of HLA class I expression can be explained by mutations on the Prim gene. Different molecular mechanisms have been shown to induce HLA class I underexpression through a Prim defect. The first signs of the importance of this chain in the control of expression of class I antigens were obtained in Daudi cells, a Burkitt’s lymphoma B cell line. Daudi cells fail to express HLA antigen because of the presence of a mutation in the initiation codon, synthesizing an abnormal mRNA that induces low efficiency in its binding to ribosomes; the mutation does not involve translation (Rosa et al., 1983). The melanoma cell line FO-1 also does not express HLA class I antigens and does not acquire them after IFN-?/ treatment. This defect is caused by the lack of transcription of the B,m gene, which has a deletion involving the first exon of the 5’ flanking region and a segment of the first intron (D’Urso et al., 1991). The lack of HLA class I antigen expression by the melanoma cell line SK-MEL-33 is caused by a guanosine deletion in the Prim gene, which involves a frame shift and the introduction of a stop codon. HLA class I expression can be reconstituted by transfection o r somatic cell fusion assays (Fellous et al., 1977; Quillet et al., 1988; Wang et al., 1993). This molecular alteration represents a somatic mutation acquired during tumor progression, since it is also detected in the autologous melanoma tissue (Wang et al., 1993). In an in situ hybridization study of HLA-negative colon carcinomas, Momburg and Koch (1989) failed to detect P,m mRNA, although the genetic mechanisms for underexpression of mRNA were not established (Cabrera et al., 1991). The mechanism may be the same as that found in colorectal tumor cell lines. Bicknell et al. (1994) reported that Prim gene mutations have been correlated with P2m phenotypic expression. The loss of expression of P,m was seen in tumor cell lines homozygous for a mutation, or heterozygous for two mutations, whereas reduced expression correlated with a mutation in one allele of Prim. A hypermutable site may exist within the first exon of the Pzm gene thiat contributes to the molecular alteration detected in some tumor cell lines (D’Urso et al., 1991). This hypothesis was confirmed by sequencing of the mutations in colorectal tumor cell lines, which showed that an 8-bp C T repeat in the leader peptide sequence was especially variable (Bicknell et al., 1994). Further proof is needed to show the relevance of these mutations in the context of genomic instability, which may play an ultimately causal role in tumorigenesis (Peinado et al., 1992). In this context, it is interesting to note that the cell lines Lovo, HCT-15, and DLD-1, which contain P2m gene mutations, express the “mutator” phe-

180

FEDERICO GARRIDO E T A L .

notype. These findings suggest an interrelation between the effect of selection of Psm mutations that lead to loss of HLA class I expression and the mutator phenotype (Bicknell et al., 1994).

3 . Abnormalities in Peptide Transported Genes Transport of class I molecules to the cell surface depends on their assembly with peptides derived from intracellular sources (Elliot et al., 1991). Peptides play a crucial role in stabilizing the structure of MHC class I in the ER ('Townsend et ul.. 1989; Cerundolo et al., 1990). Alterations in peptide processing are potential targets during tumorigenesis and confer advantages in evading tumor response. The identification of antigen-processing mutant cells helped to determine the machinery necessary to process endogenously synthesized proteins. The mouse T-cell line RMA-S (Kiirr-eet al., 1986; Townsend et al., 1989) and the human B lymphoblastoid cell line LCL 721.134 (DeMars et al., 1985) are deficient in the surface expression of MHC class I antigens, although the class I genes are apparently transcribed in the normal fashion. This phenotype is consistent with the inability to ti-ansport these peptides from the cytoplasm into the secretory pathway. These cell lines present deletions within the MHC class I1 region (Hosken and Bevan, 1990; Cerundolo et al., 1990), which contains two genes belonging to a family- of related transport proteins: the TAP genes. In fact, the defect in MHC expression can be restored b y the transfection of cDNA encoding for transport protein (Spies and DeMars, 1991). Restifo et al. (1993) have identified (in sniall-cell lung carcinoma) cells deficient in antigen processing. They found low levels of mRNA for proteasome components and peptide transporters. Loss of TAP- 1 expression has also been found in a high percentage of human cervical carcinomas (Cromme Pi al., 1994; Stern et al., 1994). 4. .Wodulation of 'ZlHC Class I Expression by Virol Iufectiorz

Viruses can interfere with antigen presentation by altering the surface expression of cell membrane proteins involved in the recognition of antigens by '1- cells (Del Val et ul.. 1989). Viral infection may result in either an enhancement or a decrease in MHC class I expression. Downregulation of hlHC class I expression has been reported to occur at both the transcriptional and posttranscriptional levels, ancl this might be a general mechanism by which virus-infected or -transformed cells escape from immune surveillance (Peek et al., 1994). Several mechanisms may conti-ihite to the down-modulation of MHC

HLA CLASS I ANTIGENS IN HUMAN TUMORS

181

class 1 expression by viral infection (Smith, 1994; Rinaldo, 1994; McFadden and Kane, 1994). DNA viruses frequently affect the assembly and transport of class I molecules to the cell surface. This transport depends on the assembly in the endoplasmic reticulum (ER) with peptides generated in the cytosol (see earlier discussion). There is convincing evidence that many viruses affect this process. For instance, protein E3/19 K from adenovirus type 2 inhibits transport to the cell surface because it binds class I peptides, which then remain in this form in the ER (Anderson et al., 1985; Korner and Burgert, 1994). In Ad12-infected cells, the levels of mRNA of the transporter genes are reduced (Rotem-Yehudar et al., 1994). In cells infected with types 1 and 2 Herpes simplex virus, HLA class I fails to become sialylated and transported to the Golgi apparatus (Hill et al., 1994). Human cytomegalovirus (HCMV) also affects MHC class I cell surface expression. Initially, it was believed that HCMV affected MHC class I expression at various levels. Competition between the HCMV UL18 gene product and class I heavy chains for binding with Pam was found to be involved in MHC class I underexpression (Browne et al., 1990). HCMV gene products interfere with processing and presentaltion by blocking the transport of peptide-loaded MHC class I complex into the medial Golgi compartment; the mechanism does not involve viral protein associated with class I molecules (Del Val et al., 1992). In this connection, HCMV reduces MHC class I protein levels by interference with the stability of class I chains, inducing their rapid degradation before they reach the cis-medial Golgi apparatus (Beersma et al., 1993). Another example of how viral DNA can affect antigen processing is the finding that suppression of the expression of peptide-transported genes inhibits cell surface class I expression in Ad 12-transformed cells (Shemesh and Ehrlich, 1993; Rotem-Yehudar et al., 1994). Finally, Epstein-Barr virus (EBV) provides an excellent example of multiple viral strategies that favor the long-term survival of virus-infected cells in immunocompetent hosts (Masucci, 1993). Allele-selective downregulation of HLA class I molecules has been documented in Burkitt’s lymphoma cell lines that retain the tumor cell phenotype (Masucci et ul., 1987). Selective allelic defects may be partly dependent on defects in antigen processing (De Campos-Lima et al., 1993). In other cases, viral control of class I expression affects the level of clilss I H-chain mRNA. Rous sarcoma virus induces the inhibition of mRNA expression in human fibroblasts (Gogusev et al., 1988). Downregulation of class I expression in tumors transformed by Ad12 occurs both at both the post-transcriptional and transcriptional levels (Schrier et al., 1983; Ackrill and Blair, 1988). The EIA gene product represses the

182

FEDERICO GARRIDO E T A L .

accumulation of MHC class I niRNA (Frieedman and Ricciardi, 1988). T h e E1A oncogene of Ad12 mediates trans-repression of MHC class I transcription by repressing the MHC class I enhancer (Ge et al., 1994). T h e R2 region within the class I enhancer acts as a negative element in Adl2-transformed cells and exhibits a stronger binding activity for COUP-TF factors that act as a transcriptional repressor (Liu et al., 1994). MHC class I down-regulation in HIV-infected cells is due to the inhibition of the MHC class I promotor by tat protein (Howcroft et al., 1993). Viral infection results in the enhancement of MHC antigen expression as a consequence of either directly by interaction with viral gene products o r indirectly by virus-induced factors released by infected cells. Cytomegalovirus infection results in the specific stimulation of promoters containing MHC NF-KBcis elements (Kowalik et al., 1993). In accordance with this observation, increased activity of the HLA-A2 promotor has been seen in a line of human T cells transfected with the HCMV IE gene (Burns et al., 1993). This may represent one of'the mechanisms by which HCMV up-regulates MHC class I expression. T h e tax protein of human T-lymphotropic virus type I induces direct transactivation of MHC class I gene expression in glial cells (Sawada et al., 1990). However, in HTLV-I-infected neuroblastoma cells, HLA class I expression was enhanced by the cytokines present in the culture supernatant and was not dependent on HTLV-I infection (Lehky et al., 1994).

E. W H Y Do MHC LOSSESOCCUR? T h e many different mechanisms known to produce alterations in HLA class I gene expression suggest that a variety of cellular genes may be the target for such alterations. It is also true that external agents like oncogenic viruses can down-regulate HLA antigens in order to escape T-cell recognition (Schrier et al., 1983). Whatever the mechanism of total, selective, or allelic HLA losses, they all lead to the absence of the restriction element required for T-cell attack. In this context, it can be posited that the hundreds of thousands of mutations detected by the random primer PCR technique, used to compare DNA from tumor and autologous lymphocytes (Peinado et al., 1992; Ionov et al., 1993), may also affect the HLA genes themselves or other genes involved in HLA regulation, assembly, or transport to the cell membrane. It is thus possible that these random mutations occurring throughout the genome could give rise to different mechanisms responsible for the downregulation of HLA class I molecules. At a particular stage of tumor development, tumor cells will escape T-cell attack by avoiding T-cell

HLA CLASS I ANTIGENS I N HUMAN TUMORS

1a3

recognition thanks to the absence of a particular HLA molecule required for antigen presentation (Garrido and Ruiz-Cabello, 1991). At this point, another immune effector mechanism, the NK cells, may come into play due to the susceptibility of the HLA-negative target (Fig. 5). It would be interesting to determine how HLA-C alleles are expressed in tumor targets, in view of the role HLA-C might play in NK alloreactivity (Ciccone et al., 1990; Colonna et al., 1993). In summary, immune selection may play a role by selecting randomly produced tumor cell clones that have lost the crucial HLA molecule required in a particular individual to present the corresponding peptide that behaves as a tumor antigen and to give rise to HLA-defective tumor targets. NK cells may represent a second line of defense able to cope with these modified targets.

T

T

HLA

+++

NK

NK

FIG. 5. Two lines of defense against tumor progression. On the left, an initially HLA+ tumor cell presents antigens that are recognized by T cells. After down-modulation of HLA molecules (right),the tumor cell becomes resistant to T-cell attack, but susceptible to lysis by N K cells. Therefore, N K cells may constitute a second line of defense in the natural history of tumor development.

FEDERICO GARRIDO E 7 Al.

184

VI. Clinical Implications .A. HI,A-I\;EGATIvE TUMORS .4ND

PROGNOSIS

It has been shown that a reduction in HLA class I antigen expression correlates significantly with poor tumor differentiation in larynx, breast, lung, and basal cell carcinomas (Redondo et al., 199lb; Cabrera et al., 1992; Esteban et al., 1990; Concha et ul., 1991a). The relationship between the loss of hlHC class I antigen expression and a decreased degree of tumor differentiation supports the assumption that the loss of hlHC class I antigen can worsen the prognosis. We have shown in larynx and breast carcinomas that tumors lacking class I expression are more aggressive and have a worse prognosis (Concha et ul., l9Ylb). Melanomas showing HLA class I losses are more invasive and have a greater metastatic capacity (Lopez Nevot et ul., 1988). Moreover, small-cell lung HLA class I-negative tumors progress and metastatize earlier (Doyle et al., 1985). In cervical cancer, early stage carcinomas have a worse prognosis if they show altered MHC class I expression (Connor et al., 1993). However, it must be clearly stated that the association between HLA class I loss and tunior invasiveness and differentiation is still an open question, as contradictory results have been obtained in tumors originating from different tissues (Gutierrez P t al., 1987; M’intzer et ul., 1990).

B. HLA

(;EKES

IN

HLX ACTIVEIMMUNOTHERAPY

I t has been shown that an HLA-\-B7gene can be transferred directly into HL.4-BS-negative melanoma patients without toxic effects (Nabel et al., 1993). DNA liposonie complexes were injected locally into the tumor lesions, and the recombinant HLA-B7 protein was detected in the tumor biopsy tissue. This gene transfer also generated an increase in the frequency of anti-HLA-B7 C‘IL precursors. In one patient, total remission of the melanoma tumor ivas obtained. This pioneering work raises the possibilit); of designing protocols for patients in whom tumors with total or partial losses of HLA class I molecules have been detected. VII. Conclusions Laboratories involved i n HLA research have been and continue to be devoted, in most hospitals, to organ transplantation and the study of MHC-associated diseases. However, it s e e m likely that the deterniination of the tunior HLA phenot),pe could become an important clinical datum in potential gene therapy (Nabel et ul., 1993). Progress in HLA class I gene transfer in cancer patients will require the precise identifica-

HLA CLASS I ANTIGENS I N HUM A N TUMORS

185

tion of the HLA antigen losses and mechanisms responsible for HLA down-regulation. As summarized in this chapter, a variety of such mechanisms have already been identified, some of which are especially frequent in certain tumors. However, routine methods for the straightforward and accurate identification o’f HLA antigen losses are not yet available. T h e XI1 HLA Histocompatibility Workshop has created a new component designated “HLA and Cancer”, which will coordinate data from different laboratories to help achieve these aims. These procedures are still a long way from being widespread in clinical practice, but there is no doubt that MHC and tumors will travel together in the years to come (Garrido and Klein, 1991). ACKNOWLEDGMENTS We thank Ms. Rosa FernAndez, Ms. Karen Shashok (member CBE, EASE), and Winston Verbi for assistance in the preparation of this manuscript. This work was supported by the Fondo de Investigaciones Sanitarias (FIS), Comisi6n Interministerial de Ciencia y Tecnologia (CICYT), Spain, the European Commission’s Biomed I Program (Proposal No. PL.931193), and Plan Andaluz d e Investigacion (PAI).

REFERENCES Ackrill, A. M., and Blair, E. (1988). Oncogene 3, 483-487. Allberti, S., and Herzenberg, L. A. (1988). Proc. Natl. Acad. Sci. USA 85, 8391-8399. Algarra, I., Ohlen, C., Pbrez, M., Ljunggren, H. G., Klein, G., and Garrido, F. (1989). 1nt.J. Cancer 44, 675-680. Andersson, M., Paabo, S., Nilsson, T., and Peterson, P. A. (1985). Cell 43, 215-222. Aramburu, J., Balboa, M. A., Izquierdo, M., and Lopez Botet, M. (1991).]. Immunol. 147, 714-721. Azuma, M., Cayabyab, M., Phillips, J. H., and Lanier, L. L. (1993).J. Immunol. 150, 20912101. Baldwin, A. S., and Sharp, P. A. (1987). Mol. Cell. Biol. 7, 305-317. Barber, L. D., and Parham, P. (1993). Annu. Rev. Cell. Biol. 9, 163-206. Barnstable, C. J., Bodmer, W. F., Brown, G., Glafre, G., Milstein, C., Williams, A. F., and Ziegler, A. (1978). Cell 14, 2-9. Barzilay, J., Kushtai, G., Plaskin, D., Feldman, M., and Eisenbach, L. (1987). Leukemia 1, 198-204. Beersrna, M. F., Bijimakers, M. J., and Ploegh, H. L. (1993).J. Immunol. 151, 4455-4464. Belich, M. P., Glynne, R. J., Senger, G., Sheer, D., and Trowsdale, J. (1994). Curr. Biol. 4, 769-776. Bernards, R. (1991). E M B O J . 10, 1119-1125. Bernards, R.,Dessain, S. K., and Weinberg, R. A. (1986). Cell 47,667-674. Bcuerle, P. A. (1991). Bzochim. Biqbhys. Acta 1072, 63-80. Biscknell, D. C., Rowan, A., and Bodmer, W. F. (1994).Proc. Natl. Acad. Sci. USA 91,47514755. Bjorkman, P. J., and Parham, P. (1990). Annu. Rev. Bzochem. 59, 253-288.

186

FEDERICO GARRIDO ET AL.

Bjorkman, P. J., Saper, M. A., Samraoui, B., Bennett, W. S., Strominger, J. L., and Wiley, D. C. (1987). Nature (London) 329, 506-5 12. Blanchet, 0..Bourge, J. F., Zinszer, H., Tatari, 2.. Degos, L.. & Paul, P. (1991).Int.J. Cancer (Su#l.) 6, 138- 145. Blanchet, 0..Bourge, J. F., Zinszner, H., Israel, A., Kourilsky, P., Dausset, J., Degos, L., and Paul, P. (1992). €'roc. Nad. Acad. Sci. USA 89, 3488-3493. Bonal, F. J., Pareja. E., Martin. J., Romero, C., and Garrido, F. (1986).J. Immunogenet. 13, 179- 186. Boon, T. (1983). Adz,. Cancer Res. 39, 121-151. Boon, T., and Kellerman, 0. (1977). Proc. Nud. Acad. Sci. USA 74, 272-275. Boucraut, J., Guillaudeux, I-., Alizadeh, hl., Boretto, J., Chimini, G., Malecaze, F., Semana, (;., Faucher, R., Pontarotti, T., and Le Bouteiller, P. (1993). Inmunogenetics 38, 117130. Brodsky. F. >I., Parham, P., Barnstable, C. J., Crumpton, M.J., and Bodmer, W. F. (1979). Immunol. Rezi. 47, 3-6 I . Browne, H., Smith, G., Beck, S., and Minson, T. (1990). Nature (London) 347, 770-772. Browning, M. J.. Krausa, P., Rowan, A,, Bicknell, D. C., Bodnier, J. G., and Bodmer, W. F., (1993). Ptot. Xatl. Acad. Sri.USA90, 2842-2845. Brusic, V.,Rudy. G . , and Harrison, L. C., (1994). lVuclezc Acids Res. 22, 3663-3665. Burke, P. A., Hirschfeld, S., Shirayoshi, Y.. Kasik, J., Hamada, K., Appella, E., and Ozato, K. (1989).J . Exp. M e d . 169, 1309-1321. Burns, I.. J., Waring, J. F., Reuter, J. J , , Stinski, M. F., and Ginder, G. D. (1993). Blood 81, 1558-1566. Burrone. 0. R., Kefford, R. F., Gilmore, D.. and Milstein, C . (1985). E M B O J . 4, 28552860. Cabrera, T., Concha. R.. Ruiz-Cabello, F., and Garrido, F. (1991). Scand. J . Immunol. 34, 147- 152. Cabrera, T., ( h r r i d o , V., Concha, A., Martin, J., Esquivias, J., Oliva, M. R., Ruiz-Cabeilo, F., Serrano, S., and Garrido, F. (1992). Immunobiology 185, 440-452. Campbell, R. D., and Trowsdale, J. (1993). Imnwnol. Today 14, 349-352. Carrefio, B. M., Winter, C. C., Taurog, J . D., Hansen, 'r. H., and Biddison, W. E. (1993). Int. I m m U t l f f l . 5, 353-360. Cella, M . , Longo, A,. Ferrara, G. B., Strominger, J. L., and Colonna, M. (1994).J. Exp. Med. 180,1235- 1242. Cerundolo, V., Alexander, J., Anderson, K., Lamb, C., Cresswell, A., McMichael, A.? Gotch, F., and Townsend, A . (1990). i\'atut-e (London) 345, 449-452. Chang, C. C., Hammer, J., Loh, J. E., Fodoi-, W.L., and Flavell, R. A. (1992). Immuiiogenetics 35,378-384. Chen, W., Khilko, S., Fecondo, J., Margulies, D. H., and McCluskey, J. (1994).J. Exp. Med. 180, 1451-1483. Christnick, E. R., Luscher, M. A,, Barber, B. H., and Williams, D. B. (1991).Nature(London) 352, (57-70. Ciccone, E., Colonna. M., Viale, O., Pende, D., Di Donato, C., Reinharz, D., Amoroso, A,, Jeannet. M., Guardiola, J.. Moretta, A, Spies, T., Strominger, J.. and Moretta, L. (1990). Proc. ,Vatl. .4cad. Sci. L'SA 87, 9794-9797. Claverie, 1. M., and Kourilsk); P. (1986). Ann. Inst. Pateur 1370, 425-442. (:levers, H . , Alarcon, B., Wileman, T..and Terhorst, C . (1988).Annu. Rev. Immu~iol.6,629662. Collins, T., Laoierre. L. A , , Fiers, kV., Strominger, J. L., a n d Poher, J. S. (1986).Proc. Natl. .4cad. Sri. L'SA 83, 446-450.

HLA CLASS I ANTIGENS IN HUMAN TUMORS

187

Collins, T. L., Kassner, P. D., Bierrer, B. E., and Burakoff, S. J. (1994). C u m Opin. Immunol. 6,385-393. Colonna, M., Bresnaham, M., Bahram, S., Strominger, J. L., and Spies, T. (1992). Proc. Natl. Acad. Sci. USA 89, 3932-3936. Colonna, M., Borselino, G., Falco, M., Ferrara, G. B., and Strominger, J. L. (1993). Proc. Natl. Acad. Sci. USA 90, 12000-12004. Concha, A., Cabrera, T., Ruiz-Cabello, F., and Garrido, F. (1991a).M . J . Cancer (Suppl.)6, 146- 154. Concha, A., Esteban, F., Cabrera, T., Ruiz-Cabello, F., and Garrido, F. (199lb). Semin. Cancer Biol., 47-54. Connor, M., Davidson, S. E., Stern, P. L., Arrand, J. R., and West, C. M. L. (1993). Int. 1. Gynecol. Cancer 32, 103-109. Cromme, F. V., Airey, J., Heemels, M. T., Ploegh, H. L., Ketaing, P. J., Stern, P. L., Meijer, C. J. L. M., and Walboomers, J. M. M. (1994).J. Ewp. Med. 179, 335-340. Csiba, A., Whitwell, H. L., and Moore, M. (1984). Br. J. Cancer 50, 699-709. Daar, A. S., Fuggle, S. V., Fabre, J. W., Ting, A., and Morris, P. J. (1984).Transplantation 38, 287-292. Darnell, J. E., Kerr, I. M., and Stark, G. R. (1994). Science 264, 1415-1421. Dausset, J. (1958). Acta Haematol. (Basel). 20, 156-160. David-Watine, B., Israel, A., and Kourilsky, P. (1990). Immunol. T o h j 11, 286-292. Davis, M. M., and Bjorkman, P. J. (1988). Nature (London) 334, 395-402. De Baetselier, P., Katzav, S., Gorelik, E., Feldman, M., and Segal, S. (1980).Nature (London) 288, 179-181. De Carnpos-Lima, P. O., Torsteinsdottir, S., Klein, G., Solitzeanu, D., and Massuci, M. G. (1993). Int. J. CanceT 53, 856-862. Del Val, M., Munch, K., Reddehase, M. J., and Koszinowski, U. H. (1989).Cell 58,305-315. Del Val, M., Hengel, H., Hacker, H., Hartlaub, U., Ruppert, T., Lucin, P., and Koszinowski, U. H. (1992).J.Exp. Med. 176, 729-738. DeMars, R., Rudrsdorf, R., Chang, C., Petersen, J., Strandtmann,J., Korn, N., Sidwell, B., and Orr, T. (1985). Proc. Natl. Acad. Sci. USA 82, 8183-8187. Dey, A,, Thornton, A,, Lonergan, M., Weissman, S., Chamberlain, J., and Ozato, K. (1992). Mol. Cell. Biol. 12, 3590-3599. Djiaballah, H., Rowe, A. J., Harding, S. E., and Rivett, A. J. (1993). Biochem. J. 292, 857862. Doyle, A., Martin, W. J., Funa, K., Gazdar, A., Carney, D., Martin, S. E., Linnoila, L., Cuttita, F., Mulshine, J., Bunn, P.,and Minna, J. (1985).J. Exp. Med. 161, 1135-1151. Driscoll, J., Brown, M. G., Finley, D., and Monaco, J. J. (1993). Nature (London) 365, 262264. D'Urso, C. M., Wang, Z., Cao, Y., Tatake, R., Zeff, R. A,, and Ferrone, S. (1991).J. Clin. Invest. 87, 284-292. Elliot, B. E., Carlow, D. A,, Rodricks, A. M., and Wade, A. (1989).Adv. CancerRes. 53, 181245. Elliot, T., Cerundolo, V., Elvin, J., and Townsend, A. (1991). Nature (London). 351, 402406. Elliot, T., Elvin, J., Cerundolo, V., Allen, H., and Townsend, A. (1992).Eur.J. Immunol. 22, 2085-2091. Engelhard, V. H. (1994a). C u m Opin. Immunol. 6, 13-23. Engelhard, V. H. (1994b). Annu. Rev. Immunol. 12, 181-207. Esteban, F., Concha, A., Huelin, C., Perez-Ayala, M., Pedrinaci, S., Ruiz-Cabello, F., and Garrido, F. (1989). Int. J. Cancer 43, 436-442.

188

FEDERICO GARRIDO E7‘ A L .

Esteban, F., Concha, A , , Delgado, M., Ruiz-Cabello, F., and Garrido, F. (1990).Br.J. Cancer 62, 1047-1051. Falk, K.,Rijtzschle, O., Stevanovic, S.. Jung, C . , and Rammensee, H. G. (1991). Nature (London) 351, 290-296. Fearon, E. R., and Volgestein, B. (1990). Cell 61, 759-767. Feldweg, A . M., and Srivastava, P. K. (1993).J. Cell. Biochem. 170, 108-115. Fellous, M.,Kamoun. M., Wiels, J., Dausset, J., Clements, G., Zeuthen, J . , and Klein, G. (1977). IrnrnunogenetirJ 5, 423-436. Fellous, M., Nir. Y., Wallach, D., Merlin, D., Rubistein, M., and Revel, M. (1982). Proc. Natl. A c a d Scz. USA 79, 3082-3090. Feltner, D. E.. Cooper, hi., Weber, J., Israel, M. A, and Thiele, C. J. (1989).J.Immunol. 143, 4292-4299. Fernandez, J. E., Concha, A., Aranegd, A.. Ruiz-Cabello, F.. Cabrera, T., and Garrido, F. (199 I). Inrmunobzolo,g 182, 440-448. Ferrcin, A . , Perez-Ayala, M., Concha, A , , Cabrera, T., Redondo, M., Oliva, M. R., RuizCabello. F., and Garrido, F. (1989).J . INtinuiiogenet. 16, 413-423. Festenstein. H. (1987). Br. M e d . Bull. 43, 217-225. , (1981). Immunol. Rezr. 60, 85-127. Festenstein, H., a n d ~ s c h m i d tW. Festenstein, H., and Garrido, F. (1986). Nuture (London) 322, 502-503. Festenstein, H., Garrido, F., Matossian, A , , Schirrmacher, V., Schmidt, W., Giorgi, L., Biasi, G., and Jardquemada, D. (1977). North Holland Biorhem., 53-67. Freeman, G. J., Gribben, J. G., Boussiotis, V. A., Kestivo, Ng, J. W., Lombard, V. A,, Jr., Gray, G.S., and Nadler, L. M.(1993). Science 262, 909-91 1. Frieedman, D. J.. and Ricciardi, R. P. (1988). Virology 165, 303-310. Fujita, ?I, Nolan, G.P., Gosh, S., and Baltimore, D. (1992). Geiirs Dev. 6, 775-787. Gabathuler, R.. Reid, G . , Kolaitis, G., Driscoll, J., and Jefferies, W. A. (1994).J.Exp. Med. 180, 14 1.5- 1425. Gaforio, J. J., Perez, M.,Algarra, I., Mialdea, M. J.. Ljunggren, H. G., a n d Garrido, F. ( 1992). I n t . J . Cancer 49, 906-9 10. Garrett, T. P. J.. Saper, M . A . , Bjorknian, 1’. J., Strominger, J. L., and Wiley, D. C. (1989). Nutiire (London) 342, 692-696. Garrido, F. (1991). 1 r i t . J . Canrer (Suppl.) 6, 1. Garrido, F.. and Festenstein, H. (1976). Folia Biol. (Pragn) 65-66. Garrido, F.. and Klein, E. (1991). Semin. Carmr B i d . 2, 1-2. Garritio, F., and Ruiz-Cabello, F. (1991). Sumin. Cancer Biol. 2, 3-10. Garrido, F.. Festenstein, H., and Schirrmacher, V. (1976). ”Vature (London) 261, 705. Garrido. F.. Perez, hi., and Torres, $1. (1979).J. Iminuiiogenet. 6, 83-86. Garrido, F., Cabrera, T., Concha, A., Clew. S., Ruiz-Cabello, F., and Stern, P. (1993). Immuriol. Today 14, 491-499. Ge, K.,Liu, X., and Ricciardi, R. P. (1994).Virology 203, 389-392. Gedde-Dahl. T.. Fossum, B., Eriksen,J. A., Thorsby, E., and Gaudernack, G. (1993).Eur.J. I l l l R l ~ ? l ( J 23, ~. 754-760. ., ( 1993). “Fundamental Immunology,” 3rd ed. (W. E. Paul, Ed.), pp. 629676. Raven Press. Germain, R. N..and Margulies, D. H. (1993). Annu. Reii. Immunol. 11, 403-450. Ghosh, S., Gifford, A . hi.. Rivere, I.. R., Tempest, P., Nolan, G. P., and Baltimore, D. (1990).Cell 62, 1019-1029. Girdlestone, J., Isaniat, M., Gewert, D., and Milstein, C . (1993). Pro(. Null. Acad. Sri. U S A 90, 11568-1 1572. Glasebrook, .4. L., and Fitch. F. \V. (1979). A’aturp (Lorrdoii) 278, 171-173.

HLA CLASS I ANTIGENS IN HUMAN TUMORS

189

Glynne, R., Powis, S. H., Beck, S., Kelly, A., Kerr, L. A., and Trowsdale, J. (1991). Nature (London) 363, 357-360. Gogusev, J., Teutsch, B., Morin, M. T., Mongiat, F., Haguenau, F., Suskind, G., and Rabotti, G. F. (1988). Proc. Natl. Acad. Sci. USA 85, 203-209. Gopas, J., Rager-Zisman, B., Bar-Eli, M., Hammerling, G. J., and Segal, S. (1989). Adv. Cancer Res. 53, 89-1 15. Gorer, P. A. (1937).J. Pathol. Bacteriol. 44,691-705. Grand, R. J., Rowe, M., Byrd, P. J., and Gallimore, P. H. (1987). Int.J. Cancer 40, 2 13-2 19. Guo, H. C., Jardetzky, T. S., Garrett, T. P. J., Lane, W. S., Strominger, J. L., and Wiley, D. C. ( 1 992). Nature (London) 360, 364-367. Gutierrez, J., L6pez Nevot, M. A,, Cabrera, T., Oliva, M. R., Esquivias, J., Ruiz-Cabello, F., and Garrido, F. (1987). Exp. Clin. Immunogenet. 4, 144-152. Gutierrez, J., Ruiz-Cabello, F., L6pez Nevot, M. A., Cabrera, T., Esquivias, J., and Garrido, F. (1990). Immunobiology 180, 138-148. Hakem, P. F., and Madrenas, J. (1990). Transplantation 50, 725-738. Hakem, R., Jezo-Bremond, A,, Le Boutelier, P., and Harper, K. (1991). Int. J . Cancer (Suppl.) 6, 2-9. Halloran, P. F., and Madrenas, J. (1990). Transplantation 50, 725-738. Hlamada, K., Gleason, S. L., Levi, B. Z., Hirschfeld, S., Appella, E., and Ozato, K. (1989). Proc. Natl. Acad. Sci. USA 86, 8289-8293. Hlarada, H., Fujita, T., Miyamoto, M., Kimura, Y., Murayama, M., Fumiyata, T., and Taniguichi, T. (1989). Cell 58, 729-739. HIarris, W., and Gill, T. (1986). Transplantation 42, 109-113. H:ass, G. G., and Nahhas, F. (1986). Am. J. Reprod. Immunol. Microbiol. 10, 39-46. HIenderson, R. A., Michel, H., Sakaguchi, K., Shabanowitz, J., Appella, E., Hunt, D. F., and Engelhard, V. H. (1992). Science 255, 1264-1266. Hlenseling, U., Schmidt, W., Scholer, H. R., Gruss, P., and Hatzopoulos, A. K. (1990).Mol. Cell. Biol. 10, 4100-4109. Hliggins, C. F. (1992). Annu. Rev. Cell. Biol. 8, 67-1 13. Hlill, V. S., Elvin, J., Willis, A. C., Aidoo, M., Allsopp, C. E. M., Gotch, F. M., Gao, X. M., Takiguchi, M., Greenwood, B. M., Townsend, A. R. M., McMichael, A. J., and Whittle, H. C. (1992). Nature (London) 360, 434-439. Hill, A. B., Barnett, B. C., McMichael, A. J., and McGeoch, D. J. (1994).J. Immunol. 152, 2736-2741. Hochstenbach, F., David, V., Watkins, C., and Brener, M. B. (1992). Proc. Natl. Acad. Sci. USA 89,4734-4738. Hosken, N. A., and Bevan, M. J. (1990). Science 248, 367-370. H[owcroft, T. K., Richarson, J. C., and Singer, D. S. (1993). E M B O J . 12, 3163-3169. Hhczko, E. L., Bodnar, W. M., Benjamin, D., Sakaguchi, K., Zhu, N. Z., Shabanowitz, J., Henderson, R. A,, Appella, E., Hunt, D. F., and Engelhard, V. H. (1993).J. Immunol. 151,2572-2588. Hhi, K., Grosveld, F., and Festenstein, H. (1984). Nature (London) 311, 750-752. Hunt, J. S., and Hsi, B. L. (1990).Am. J. Reprod. Immunol. 23, 57-60. HLunt, D. F., Henderson, R. A., Shabanowitz, J., Sakaguchi, K., Michel, H., Sevilir, N., Cox, A., Appella, E., and Engelhard, V. H. (1992). Science 255, 1261-1263. Hwang, C., Sinskey, A,, and Lodish, H. F. (1992). Science 257, 1496-1502. Ionov, Y., Peinado, M. A., Malkhosyan, S., Shibata, D., and Perucho, M. (1993). Nature (London) 353, 558-561. Iwael, A., Kimura; A , , Fournier, A,, Fellous, M., and Kourilsky, P. (1986). Nature (London) 322, 743-746.

190

FEDERICO GARRIDO E T A L .

Jardetzky, T. S., Lane, W. S.,Robinson, R. A,, Madden, D. R., and Wiley, D. C. (1991). Naturp (London) 353, 326-329. Karjalainen, K. (1994). CUT. Opzn. Immunol. 6, 9-12. Karre, K., Ljiinggren, H. G., Piontek, G., and Kiessling, R. (1986).Nature (London) 319, 675-678. Katzav, S., Segal, S., and Feldrnan, M. (1985).J. Natl. Cancer Inst. 75, 307-317. Kavathas, P., Bach. F., and DeMars, R. (1980). Proc. Natl. Acad. Sci. USA 77, 4251-4255. Keating, P. J . , Crornme. F. V., Duggan-Keen, M., Snijders, P. J. F., Walboorners, J. M. M., Hunter, R., Dyer, P. A., and Stern, P. L. (1995). 8 r . J . Cancer, in press. Kelly, A,, Powis, S. H., Glynne, R., Radley. E., Beck, S., and Trowsdale, J. (1991). Nature (London) 353, 667-668. Khilko, S. N., Corr, M., Boyd, L. F., Lees, A., Inrnan, J. K., and Margulies, D. H. (1993).J. Biol. Chem. 268, 15425- 15434. Kieran, M., Blank, V., Logeat, F., Vandekerckhove, J.. Lottspeich, F., Le Bail, 0..Urban, M. B., Kourilsky, P., and Beuerle, P. A. (1990).Cell 62, 1007-1018. Kirnura, A., Israel, A., Le Bail, O., and Kourilsky, P. (1986). Cell 44, 261-272. Klein, G . , Sjogren, H., Klein, E., and Hellstroni, K. E. (1960). Cancer Res. 20, 1561-1572. Kopp, F., Dahlrnan, B., and Hendil, K. B. (1993).J.Mol. Biol. 229, 14-19. Korner, H., and Burgert, H. G. (lY94).J. Vzrol. 68, 1442-1448. Kowalik. T. F., Wing, B., Haskill, J. S., Azizkhan, J. C., and Baldwin, A. S., Jr. (1993).Proc. iVatl. Acud. Scz. USA 90, 1107- 1 1 11. Kronenberg, M., Siu, G., Hood, L. E., and Shastri, N. (1986).Annu. Rev. Immunol. 4, 529591. Kushtai, G., Barzilay, J.. Feldrnan, M., and Eisenbach, L. (1988). Oncogene 2, 119-125. Lehky, T. J.. (;owan, E. P., Lanipson, L. A., and Jacobson, S. (1994).J.Virol. 68, 1854-1863. Li, Z., and Srivastava, P. K. (1993). E M B O J . 12, 3143-3151. Litwin, V., Guniperz, J., Parharn, P., Phillips, J. H., and Lanier, L. L. (1994).J . Exp. Med. 180, .537-543. Liu, X., Ge, R., Westmoreland, S.,Cooney, A. J., Tsai, S. Y., Tsai, M. J., and Ricciardi, R. P. (1994).Oncogene 9, 2183-2190. Ljunggren, H. G., and Karre, K. (1990).Immunol. Todq 11, 237-244. Ljunggren, H. G . , Starn, N. S., Ohlen, C., Neefjes, J. J., Hoglund, P., Heemels, M. T., Bastin, J.. Schurnacher, T. N. hi., Townsend, A., Karre, K., and Ploegh, H. L. (1990). Nature (London) 346, 476-480. Lopez Nevot, M. A., Garcia, E., Pareja, E.. Bonal, F. J . , Martin, J . , Ruiz-Cabello, F., and Garrido, F. (1986).J. Inmunogewf. 13, 219-227. Lopez Nevot, M. A., Rornero, C., Ruiz-Cabello, F., Pedrinaci, S., Gutierrez, J., Oliva, M. R., Huelin, C., and Garrido, F. (1987).Histocum. Testing X . lnt. U'orkJhop, 473. Lopez Nevot, $1. A, Garcia, E., Rornero, C., Oliva, M . R., Serrano, S., and Garrido, F. (1988). Exp. Clin. Inimunogenet. 5 , 203-2 12. L6pez Nevot, M. A . , Esteban, F., Ferron, A . , Gutierrez, J., Oliva, M. R., Rornero, C., Huelin, C., Kuiz-Cabello, F., and Garrido, F. (1989).B r . J . Cancer 59, 221-226. Lozano, F.. Santos-Aguado, J., Borche, L., Places, L., Dornenech, N.,Gaya, A,, Vilella, R., and Vives, J . (1989). Inimunogen~tics30, 50-53. Lu, Y. Y., Blair, D. G . , Segal, S., Shih. T. Y., and Clanton, D. J. (1991). lnt. J. Cancer 6 (Suppl.). 4.5-5J. Madden, D. R., Gorga, J. C., Strorninger, J. L., and Wiley, D. C . (1992). Cell 70, 1035-1048. Marincola, F. M., Shamamian, P., Alexander, R. B., Gnarra, J. R., Turetskaya, R. L., Nedospasov, S. A., Simonis, T. B., Taubenberger, J. K., Yannelli, J.. Mixon, A , , Restifo, N. P.. Herlyn, M.,and Kosenberg, S. A. (l994).J.Ininiunol. 153, 1225-1237.

HLA CLASS I ANTIGENS IN HUMAN TUMORS

191

Marks, M., Hallenbeck, P.. Nagata. T., Segars, J., Appella, E., Nikodem, V., and Ozato, H. (1992). E M B O J . 11, 1419-1435. Maryanski, J. L., Romero, P., Van Pel, A., Boon, T., Salemme, F. R.,Cerottini, J. C., and Corradin, G. (1991). Int. Immunol.3, 1035-1042. Maschek, U., Pulm, W., Segal, S., and Hammerling, G. J. (1989). Mol. Cell. Biol. 9, 31363142. Masucci, M. G. (1993). Cum. 0P;n. Immunul. 5, 693-700. Masucci, M. G., Torsteindottir, S., Colombani, J., Brautbar, C., Klein, E., and Klein, G. (1987). Proc. Natl. Acad. Sci. USA 84, 4567-4571. Matsumura, M., Fremont, D. H., Peterson, P. A., and Wilson, I. A. (1992). Science 457, 927-934. McFadden, G . , and Kane, K. (1994). Adv. Cancer R a . 63, 117-209. Melief, C. J. M. (1992). In "Advances in Cancer Research" (G. Klein and G. Van de Woude, Eds.), Vol. 58, pp. 143-175. Academic Press, San Diego, CA. Melief, C. J. M., and Kast, W.N. (1991). Semin. Cancer Biol. 2, 347-354. Michalek, M. T., Grant, E. P.,Gramm, C., Goldberg, A. L., and Rock, K. L. (1993). Nature (London) 363,552-554. Miyamoto, M., Fujita, T., Kimura, Y., Murayama, M., Harada, H., Sudo, Y., Miyata, T., and Taniguichi, T. (1988). Cell 54, 903-913. Mizoguchi, H., OShea, J., Longo, D., Loeffer, C., McVicar, D., and Ochoa, A. (1992). Science 258, 1795- 1798. Moller, P., and Hammerling, G. (1992). Cancer Suruqs 13, 101-128. Moller, P., Koretz, K., Schlag, P., and Momburg, F. (1991). I d . J . Cancer (Suppl.) 6 , 155162. Momburg, F., and Koch, S. (1989).J. Exp. Med. 168, 309-314. Momburg, F., Degener, T., Bacchus, E., Moldenhauer, G., Hammerling, G. J., and Moller, P. (1986). Int. J. Cancer 37, 179- 184. Momburg, F., Neefjes, J. J., and Hammerling, G. J. (1994). Cum. Opn. Immunol. 6, 32-37. Monaco, J. J. (1992). Immunol.Today 13, 173-179. Moretta, L., Ciccione, E., Mingari, M. C., Biassoni, R., and Moretta, A. (1994a). Adv. Immunol. 55, 341-380. Moretta, A., Vitale, M., Sivori, S., Bottino, G., Morelli, L., Augugliaro, R., Barbaresi, M., Pende, D., Ciccone, E..L6pez Botet, M., and Moretta, L. (1994b).J. Exp. Med. 180,545555. Miiller, M., Briscoe, J., Laxton, C., Guschin, D., Ziemiecki, A., Silvennoinen. O., Harpur, A. G., Barbieri, G., Witthuhn, B. A., Schindler, C., kllegrini, S., Wilks, A. F., Ihle, J. N., Stark, G., and Kerr, I. M. (1993). Nature (Lotldon) 366, 129-135. Nabel, G. J., Nabel, E., Yang, Z. Y., Bernard, A. F., Plautz, G. E., Gao, X.,Huang, L., Shu, S., and Gordon, D. (1993). Proc. Nail. Acad. Sci. USA 90, 11307-11311. Natali, P. G., Bigotti, A,, Nicotra, M. R.,Viora, M., Manfredi, D., and Ferrone, S. (1984). Cancer Res. 44,4679-4680. Neefjes, J. J., Hammerling, G. J., and Momburg, F. (1993).J. axp. Med. 178, 1971-1983. Nolan, G. P., Gosh, S., Liou, H. C., Tempst, P., and Baltimore, D. (1991). Cell 64, 961969. Nossal, G. J. V. (1994). Cell 76, 229-239. Oliva, M. R.,Cabrera, T., Esquivias, J., Perez-Ayala, M.. Redondo, M., Ruiz-Cabello, F., and Garrido, F. (1990). Int. J. Cancer. 46, 426-431. Rmer, E. G., Harty, J. T., and Bevan, M. J. (1992). Nature (London)353, 852-855. Parham, P.,and Barber, L. D. (1994).J. Exp. Med. 180, 1191-1194. Parham, P., Lawlor, D. A., Salter, R. D., Lomen, C. E., Bjorkman, P. J., and Ennis, P. D.

192

FEDERICO CARRIDO ET AL.

( 1989). “Immunobiology

of HLA. lmniunogenetics and Histocompatibility” (B. Dupont, Ed.), Vol. 11, pp. 10-33. Springer-Verlag. Peek, R.. Westphal, J. R., Pruijn, G. J., Van der Kemp, A. J.. and Van Venrooij, W.J. (1994). Clin.Exp. Inimunol. 96,395-402. Peinado, M. A., Malkhosyan, S., Velazquez. A.. and Perucho, M. (1992).Prw. Natl. Acad. S C ~W . A 89, 10065-10069. Pellegrino, M.A., Ferrone, S.. Reisfeld, R. A., hie, R. F., and Ckdub, S. H. (1977).Cancer

40.36-41.

pcirez, M.,Algarra, I., Ljunggren, H. G., Caballero, A., Mialdea, M.J.. Gaforio, J. J., Klein, G., Klrre, K., and Garrido, F. (1990).6 1 . J. Cancer 46,258-261. Pestka, S.,Langer, J. A., Zoon, K. C., and Samuel, C. E. (1987).Annu. Rev. Biochen. 56,

727-777.

Petersen. B. L., Petersen. C. L.,Braendstrup, O., Mouritsen, S., Engel. A. M.,Svane, I. M., and Werdelin, 0.(1993).APMIS 101,529-936. Plaksin. D., Baeuerle, P. A., and Eisenbach. L. (1993).J. Exp. Med. 177, 1651-1662. Ploegh, H. L., Cannon, L. E.,and Srominger, J. L. (1979).Proc. Natl. Acad. Sci. USA 76,

2273-2277.

Poggi. A., Sargiacomo, M.. Biassoni. R.. Pella, N., Sivori, S., Revelo, V., Costa, P., Valteri, M., Russo. G., Mimgani, hl. C.. Peschle, C., and Moretta, L. (1993).Proc. Nd.A d . Sci. I’SA 90,4463-4469.

Poltack, 34. S., Livingston, P. 0.. Fogh, J.. Carey. 1.E., Oettgen, H. F., and Dupont, S. 0. (1980).Tissue Antigens 15,249-256. Potter, D. A., Larson, C. J., Eckes, P., Schmid, R. M.,Nabel, G.J., Verdine, G. J., and Sharp, P. A. (1993).1.Bid. Chem. 268, 18882- 18890. Powis, S.J., Deverson, E. V., Coadwell, M?J.. Ciruela. A., Huskisson, N. S., Smith, H., Pujol Borrell, R.. Todd. I., and Londei. hi. (1986).1.Mol. Bid. Mtd. 3, 159-165. Powis, S. J., Deverson, E. V., Coadwell. W.J.. Cirueka, A.. Huskisson, N. S., Smith, H., Butcher, G.W., and Howard, J. C. (1992). Nature (London) 357, 211-215. Quillet. A., Presse. F., Marchiol-Fournigault, C., Harell-Bellan, A., Benbunan, M., Ploegh, H.,and Fradelizi. D. (1988).J. Immanol. 141, 17-20. Rechsteiner. M.,Hoffman, L., and Dubiel, N’. (1993).J. Biol. Cltm. 268, 6065-6068. Redman, C. 1%’G., . McMichael. A. J., Stirrats, G. M., Sunderland, C. A., and Ting, A. (1984).Ininiunolo~54,4547-4552. Redondo. M..Ruiz-Cabello. F.. Concha, A., Cabrera. T., Pkrez-Ayala, M.,Oliva, M.R., and Garrido. F. (1991a).Cancer Res. 51, 2465-2468. Redondo. M.,Concha, A.>Cueto, A.. Garrido, F., and Ruir-Cabello, F. (1991b).CuncerRes.

51,4948-4994.

Kestifo, K. P., Esquivel, F.. Kawakami, Y.,Yewdell. J.. hiul6, J. I., and Rosemberg, S. A. (I 993).J. Ex#. AJed 177, 265-272. Rinaldo. C. R. (1994). AWLJ. Pathol. 144,637-650. Roberts, J. M.,Taylor, <:. 11, Melling, G. C.. Kingsland, C. R., and Johnson, P. M. (1992). Immunolug?’75, 202-205. Rodewald, H. R., Moingedn, P., Lucich, J. L., Dosiov, C., L6pe2, P., and Reinherz, E. L. ( 1992).Cell 69, 139- 150. Rodriguez. F., Peran, F., Garrido. F., and Ruit-Cabello, F. (1994).Iwnunogenetics 39, 161-

167. Rosa, F., Berissi. H., M’eissenhdch, J.. Maroteaux, L., Fellous. M., and Revel, M. (1983). EMBO J. 2,239-243. Rosenthal, A. S..and Shevach. E. M. (1973bJ.Exp. Med. 138, 1194-1212.

HLA CLASS I ANTIGENS IN HUMAN TUMORS

193

Rotem-Yehudar, R., Shechter, H., and Ehrlich, R. (1994). Gene 144,265-270. R(%zschke,O.,Falk, K., Stevanovic, S., Jung, G., Walden, P., and Rammensee, H. G. (1991). Eur. J. Immunol. 21,2891-2894. R6tzschke, O., Falk, K., Stevanovic, S., Jung, G., and Rammensee, H. G. (1992). Eur. J . Immunol. 22, 2453-2456. Rigwland-Jones, S., Sutton, J., Ariyoshi, K., Dong, T., Gotch, F., McAdam, S., Whitby, D., Sabally, S., Gallimore, A., Corrah, T., Takiguchi, M., Schultz, T., McMichael, A,, and Whittle, H. (1995). Nature (London) 1, 59-64. Ruiter, D. J., Mattijssen, V., Broecker, E. B., and Ferrone, S. (1991). Semin. CancerBioL 2, 35-45. Ruiz-Cabello, F., Lopez Nevot, M. A,, and Garrido, F. (1987). Cancer Metastasis Rev. 233, 119-128. Ruiz-Cabello, F., Lopez Nevot, F., Gutierrez, J., Ferron, A., Esteban, F., Rivas, C., Piris, M. A,, Pedrinaci, S., Oliva, M. R., Romero, C., Huelin, C., and Garrido, F. (1989). Clin. Exp. Metastasis 77,213-226. Ruiz-Cabello, F., Perez-Ayala, M., Gomez, O., Redondo, M., Concha, A,, Cabrera, T., and Garrido, F. (1991a). Znt.J. Cancer. (Sup$.) 6, 123-130. Ruiz-Cabello, F., Klein, E., and Garrido, F. (1991b). Immunol. Lett. 29, 181-190. Ruiz-Cabello, F., Torres, M. J., Skoudy, A., Berrozpe, G., Cabrera, T., Real, F. X., and Garrido, F. (1 995). Tissue Antigens, in press. Salter, R. D., Benjamin, R. J., Wesley, P. K., Buxton, S. E., Garrett, T. P. J., Clayberger, C., Krensky, A. M., Norment, A. M., Littman, D. R., and Parham, P. (1990). Nature(London) 345,4 1-46. S,altzman, A. G., and Weinman, R. (1989). FASEB J . 3, 1723-1733. S,ambrook,J. F. (1990). Cell 61, 197-199. S.aper, M. A,, Bjorkman, P. J., and Wiley, D. C. (1991).J. Mol. Biol. 219, 277-319. Sawada, M., Suzumura, A., Yoshida, M., and Marunouchi, T. (1990). J. Virol. 64, 40024006. Schrier, P. I., and Peltenburg, L. T. (1993). Adv. Cancer Res. 60, 181-246. Smchrier, P. I., Bernard, R., Vaessen, R. T., Houweling, A., and Van Der Eb, A. J. (1983). Nature (London) 305, 771-775. Schrier, P. I., Versteeg, R., Peltenburg, L. T., Plomp, A. C., Van't Veer, L. J., and KriiseWolters, K. M. (1991). Semin. CancerBiol. 2, 73-83. Schmidt, W., Alfield, G., and Festenstein, H. (1979).J. Zmmunogenet. 8, 311-321. Schuai, K., Schindler, C., Prezioso, V. R., and Darnell (1992). Science 258, 1808-1812. Seung, S., Urban, J. L., and Schreiber, H. (1993). J. Exp. Med. 178, 933-940. Shemesh, J., and Ehrlich, R. (1993).J. Biol. Chem. 268, 15704-15710. Shindler, C., Shuai, K., Prezioso, V. R., and Darnell, J. E. (1992). Science 257, 809-813. Singer, D., and Maguire, J. (1990). Crit. Rev. Zmmunol. 10, 235-257. Slingluff, C. R., Jr., Cox, A. L., Henderson, R. A., Hunt, D. F., and Engelhard, V. H. (1993).J. Immunol. 150, 2955-2963. Slingluff, C. L., Jr., Hunt, D. F., and Engelhard, V. D. (1994). Cum. Opin. Immunol. 6,733740. Smith, M. (1991). Semin. Cancer Biol. 2, 17-21. Smith, G. L. (1994). Trendc Microbiol. 2, 81-88. Snell, G. D. (1958).J. Natl. Cancer Inst. 21, 843-850. Soong, T. W., and Hui, K. M. (1991). Int.,]. Cancer(Supp1.) 6, 131-137. Sorger, S. B., Paterson, Y., Fink, P. J., and Hedrick, S:M. (1990). J. Immunol. 144, 11271135.

194

FEDERICO GARRIDO E T AL.

Spies, T., and DeSiars, R. (1991). X a t u r e (London) 351, 323-324. Spies. T., Bresnahan, M.,Bahram, S.. Arnold, D., Blanck, G., Mellins, E., Pious, D., and DeMars, K. (1990). Sature (London) 348, 744-747. Snvastava, P. K.. L'dono, H.. Blachere, N. E.. and Li, 2. (1994). Iniluuogenetics 39,93-98. Stern. P. I.., Keiienians-P. Helmerhorst. T.J.. Stukart, hl. J., and Meijer, C. J. (1994).Br.1. Cancer 69, 1 176- 1 I8 1 . Storkus, W. J., Salter, R. D.. Alexander. J., Ward. F. E., Ruiz, K. E., Cresswell, P., and Dawson, J. R. (1991). Pror. SaB Arad. Sri. VSA 88, 5989-5992. ~ 129-142. Swain. S. (1983). Immicnol. R F Z74, Fdnaka, K.. Isselbacher. K. J., Khour); G., and Jay, G. (1985). Science 228, 26-30. Ting, J. P. Y.,and Baldwin, A. S. (1993). Cum. Opin. Immuiiol. 5, 8-16. Torres, L. M., Cabrera, T., Concha, A., Oliva. M. R., Ruiz-Cabello, F., and Garrido, F. (1993). Tissue Antigens 41,65-71. 'Iownsend, A., and Bodmer, H. (1989). Annu. Rev. Immunol. 7,601-624. Townsend, A., Rothbard, J., and Gotch, F. M. (1986). C e l l 4 4 , 959-968. Townsend, A.. Ohlen. C., Bastin. J., Ljunggren, H. C.. Foster, L., and Karre, K. (1989). Xature (London) 340,443-448. Traversari. C., van der Bruggen, P., Luescher, I. F., Lurquin, C., Chomez, P., Van Pel, A., De Plaen, E., Amar-Costesec, A.. and Boon, T. (1992).J . Exp. Med. 176, 1453-1457. 'Trinchieri, G. (1989). Adv. Immunol. 47, 187-376. Trinchieri, G. ( 1 994). J. Exp. Med. 180,417-42 1. Trowdale, J., Hanson, I., Mockridge, I.. Beck, S., Townsend, A., and Kelly, A. (1990). Naiure (London) 348, 74 1-744. Udaka, K., Tsomides, T. J., and Eisen, H. N. (1992). Cell 69,989-998. van der Bruggen, P., Traversari, C., Chomez, P., Lurquin, C., De Plaen, E.,Van den Eynde, B., Knuth, A., and Boob, T. (1991). Science 254, 1643-1647. van der Bruggen, P., Szikord, J. P., Boel, P., Wildman, C., Soniville, M., Sensi, M., and Boon, T. (1994). Eur. J. Imtnunol. 3, 2134-2140. van't Veer, L. J., Beijersbergen. R. L., and Bernards, R. (1993). EMBOJ. 12, 195-200. Versteeg, R., Noordermeer, I . A., Kriise-Wolters, M., Ruiter, D. J., and Schrier, P. I. (1988). EMBOJ. 7, 1023-1029. von Boehrner, H. (1994). Cell 76,219-228. Wallich, R., Bulbuc, N., Hammerling, G. J.. Katzav, S., Segal, S., and Feldman, M. (1985). Nature (London) 315, 301-305. Wallny, H . J., Deres, K., Faath, S.,Jung, G., Van Pel, A., Boon, T.. and Rammensee, H. G. (1992). Int. Imniunol. 4, 1085-1090. Wang. Z., Cao, Y.,Albino, A. P., Zeff, R. X.. Houghton, A., and Ferrone, S. (1993).J. Clin. Invest. 91, 684-692. Weber, S., Traunecker, A., Oliveri. F., Gerhard, W., and Karjalainen, K. (1992). Nature (London) 356, 793-795. Wei, X., and Orr. H. T. (1990). Hum. Imrnuriol. 29, 131-142. Wei, M. L., and Cresswell. P. (1992). .Vature (London) 356, 443-446. Weissman, J. D., and Singer, D. (1991). Mol. Cell. Biol. 11, 4217-4227. Wintzer, H. O., Benzing, M., and ron Kleist, S. ( 1 990). Br. J. Cancer 62, 289-295. brig, S. Y., Morishima, Y., Collins. N. H., Alton. T., Pollack, M. S., Yunis, E. J., and Dupont, B. (1984). I~nmunogenetirs19, 217-232. k n g , Y.,Waters-J. B., Friih, K., and Peterson, P. A. (1992). Proc. Nutl. Acad. Sci. USA 89, 4928-4932.

HLA CLASS I ANTIGENS IN HUMAN TUMORS

195

Yano, O., Kanellopoulus, J., Kieran, M., Le Bail, O., Israel, A., and Kourislky, P. (1987). EMBOJ. 6,3317-3324. Zeff, R.A., Nakagawa, M., Mashimo, H., Gopas, J., and Nathenson, S. G. (1990). Tramplantatzon 49, 803-808. Zinkernagel, R. M., and Doherty, P. C. (1974). Nature (London) 248, 701-702.

This Page Intentionally Left Blank

MOLECULAR EPIDEMIOLOGY OF EPSTEIN-BARR VIRUS INFECTION Jan W. Gratamal and lngemar Ernberg2 'Department of Clinical and Tumor Immunology, Daniel den Hoed Cancer Center, Rotterdam, The Netherlands, and ZMicrobiology and Tumorbiology Center, Karolinska Institute, S-I71 77 Stockholm, Sweden

I. Introduction 11. EBV Genome and Gene Expression

111. EBV Typing at the DNA Level (Genotyping) A. Introduction B. Differential Detection of Types A and B EBV C. Genomic Heterogeneity of the LMP- 1 Coding Region in the BarnHI N Fragment D. Other Restriction Fragment Length Polymorphisms E. Defective EBV Genotypes F. RFLP as a Tool to Study EBV Transmission G. Conclusions IV. EBV Typing at the Protein Level (Ebnotyping) A. Introduction B. EBV-Encoded Proteins in EBV-Transformed Cells C. Ebnotyping Studies of EBV Carrier Status and Transmission Patterns D. Conclusions V. Differential Recognition of EBV Genotypes by the Immune System A. EBV Genotype-Specific Antibody Responses B. EBV Genotype-Specific Cellular Immune Responses VI. Final Conclusions: Implications for the Biology of EBV Infection A. Epidemiology: Sorting EBV Genotypes and Ebnotypes in Order of Their Biological Significance B. Epidemiology: Viral Evolution under Immunological Pressure? C. Transmission: Support for and Challenge of Existing Concepts D. Pathogenesis: New Tricks from an Old Dog E. Pathogenesis: Lack of Effective Immune Surveillance References

I. Introduction

Epstein-Barr virus (EBV) is a y-herpes virus and consists of a core containing a double-stranded DNA molecule of 172,000 base pairs (bp), an icosahedral capsid, and an envelope enclosing the capsid. Similar to other herpes viruses, primary infection is followed by lifelong carrier status of the virus. The two major target cell types for EBV infection are the B lymphocytes, in which the infection is largely nonproductive or

-

197 ADVANCES IN CANCER RESEARCH, VOL. 67

Copyright 0 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

198

J A N W. GRATAMA A N D INGEMAR ERNBERG

latent (Nilsson et at., 1971), and stratified squamous epithelium, in which viral replication occurs (Sixbey et al., 1984; Greenspan et al., 1985). EBV enters B lymphocytes following interaction between the viral membrane glycoprotein gp340/220 and the B-cell receptor for the C3d complement fragment (CD2 1) [reviewed by Nemerow et al. (1990)l. Although CD21 can also be the port of entrance for EBV into epithelial cells (Birkenbach et af., 1992), an alternative entry route into otherwise resistant epithelial cells is the coupling of virions by EBV-specific polymeric immunoglobulin A to the secretory component, a transmembrane protein expressed by epithelial cells, followed by endocytosis (Sixbey and Yao, 1992). EBV has a worldwide distribution; -95% of adults carry the virus. Most infections occur within the first years of life and are asymptomatic. EBV infects humans via salivary contact in most nosocomial infections (Hoagland, 1955; Evans, 1960; Morgan et al., 1979; Sumaya and Ench, 1986). A second site of virus production is the urogenital tract (Sixbey et al., 1986; Israele et al., 1991; Taylor et al., 1994b). In affluent populations with high standards of hygiene, primary EBV infections may not occur until late adolescence or adulthood (Niederman et al., 1970; Henle and Henle, 1970). Approximately half of these late primary infections result in the clinical syndrome of infectious mononucleosis (IM; Henle et al., 1968).Occasionally, I M may be caused by transmission of EBV via blood transfusion (Wising, 1942; De Vos and Kuipers, 1951; Gerber et al., 1969; Blacklow et at., 1971; Turner et al., 1972; Henle and Henle, 1985) or bone marrow transplantation (BMT) (Gratama et al., 1994a). T h e concept that the oropharyngeal epithelium and not the lymphoid compartment is the main site of intermittant production of infectious virus has recently been challenged by Anagnostopoulos et al. (1995),who demonstrated in IM patients latent and, importantly, productive EBV infection in lymphoid cells on the surface of tonsillar epithelium and within tonsillar crypts, but not in any epithelial cells. In line with these observations, Tao et al. (1995) demonstrated productive EBV infection in mucosal lymphocytes but not in nasopharyngeal epithelial cells of EBV carriers without evidence of EBV-related disease. Since its discovery (Epstein et al., 1964), EBV has been associated with several human malignant tumors of either hematopoietic or epithelial origin. T h e EBV+ cases of Burkitt’s lymphoma (BL) are pediatric B-cell tumors endemic in the African malaria belt, while the EBV- BL are not restricted with respect to age and geographical distribution (Magrath, 1990). EBV is also associated with certain histological subtypes of Hodgkin’s disease [HD; reviewed by Pallesen et al. (1993)]. EBV genomes can be detected in the Reed-Sternberg cells of most cases with mixed cellular histology, in -30% of cases with nodular sclerosing histology, and in only 10% of‘ cases with lymphocytic predominance. Similarly, a wide

-

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

199

range of peripheral T-cell lymphomas may be EBV+ [reviewed by Pallesen et al. (1993)l. EBV is consistently present in the tumor cells of nasal T-cell lymphomas, whereas it appears to infect only a fraction of the tumor cells in other peripheral T-cell lymphomas. The best documented example of an EBV+ epithelial tumor is nasopharyngeal carcinoma (NPC), which is endemic in regions of southern China (Zur Hausen et al., 1970). All investigated NPC tumors carry EBV, which has not been precedented by any other virus-associated tumor (Andersson-Anvret et al., 1978). EBV has also been detected in lymphoepitheliomas occurring in other organs such as parotid gland (Krishnamurthy et al., 1987), thymus (Leyvraz et al., 1985), lung (Weiss et al., 1989), and stomach (Shibata el: al., 1991; Imai et al., 1994), as well as in a minority of cases of gastric adenocarcinoma (Shibata and Weiss, 1992; Imai et al., 1994) and breast cancer (Labrecque e t a ) . , 1995). In immunodeficient individuals it has been recognized for more than 10 years that uncontrolled proliferations of EBV-carrying B lymphocytes may occur (Hanto et al., 1982). These immunodeficiencies may be congenital, such as the X-linked lymphoproliferative disease, or acquired, such as in organ and bone marrow transplant (BMT) recipients, rheumatic patients receiving immunosuppressive therapy and in patients infected with the human immunodeficiency virus (HIV) (reviewed by Thomas et al., 1991; Kame1 et al., 1995). The lesions of oral hairy leukoplakia in the latter group of patients constitute clinically apparent foci of intense EBV replication in lingual epithelium (Greenspan et al., 1985). Recently, EBV has also been demonstrated in smooth-muscle tumors arising in immunosuppressed children, i.e., organ transplant recipients (Lee et al., 1995) or patients with AIDS (McClain et al., 1995). T h e availability and use of molecular analytical techniques at the DNA and protein levels allowed the recognition of heterogeneity between EB viruses carried by various laboratory B-cell lines that had been derived from BL tissues (Strnad et al., 1981; Hennessy et al., 1983; Hennessy and Kieff, 1983; Dambaugh el ad.,1984; Dillner et al., 1984; Sculley et al., 1984; Adldinger et al., 1985; Ernberg et al., 1986; Zimber et al., 1986). Similarly, heterogeneity between EB viruses derived from patients with IM, rheumatoid arthritis, and healthy cafriers was detected at the DNA and protein levels (Ernberg et al., 1986; Sculley et al., 1987). Until that time, insight into EBV infection patterns had only been obtained by serological studies (Henle et al., 1968; Gerber et al., 1969; Henle and Henle, 1970, 1985; Joncas and Mitnyan, 1970; Niederman et al., 1970; Wahren et al., 1970; Blacklow et al., 1971; Sawyer et al., 1971; Fleischer et al., 1981). This chapter summarizes the information obtained by the molecular characterization of EBV diversity and discusses its relevance for viral epidemiology, transmission, and pathogenesis.

200

JAN W. GRATA'MA AND INGEMAR ERNBERC;

It. EBV Genome and Gene Expression T h e prototype EBV has been derived from the B95-8 cell line, which was established by infecting marmoset B cells with EBV from an individual with IM (Miller et al., 1972). The B95-8 EBV genome has been completely sequenced (Baer et al., 1984) and extensively mapped for transcripts, promotors, open reading frames, and other structural elements [reviewed by Farrell (1989); Fig. 11. It is separated into short and long unique sections by a variable number of tandemly reiterated 3.1kb-long internal repeats (Sample et al., 1986), and it is flanked at either end by multiple tandem 0.5-kb-long terminal repeats (Raab-Traub and Flynn, 1986). Following entry into the B lymphocyte, the viral genome circularizes by covalent linkage of the terminal repeats and is then maintained as a multicopy episomal plasmid (Hurley and Thorley-Lawson, 1988). Because both the number of terminal repeat sequences at either end of the linear EBV genome and the extent of overlapping of terminal repeat sequences during episome formation are variable, each new circularization event leads to a differently sized terminal repeat fragment (Raab-Traub and Flynn, 1986). In latently infected B cells, the replication o f episonies is regulated and occurs in parallel with host cell proliferation, so that the genomic structure and copy numbers of the episomes remain stable (Yates and Guan, 1991). Hence, the size of the terminal repeat fragment represents a constant clonal marker for the episome and is informative for the clonality of EBV+ (tumor) cell populations (Raab-Traub and Flynn, 1986; Brown et al., 1988; Cleary et al., 1988). T h e viral gene products that can be expressed during latency are six nuclear proteins, termed EBNA- 1-6, and three membrane proteins, termed LMP-1, LMP-PA, and LMP-2B [see reviews by Kieff and Liebowitz ( 1990) and Rogers et al. ( 1992)]. An alternative nomenclature for some of these proteins is being used by these and other groups, in which the BurnHI E-encoded proteins EBNA-3, -4, and -6 are referred to as EBNA-h, -3b, and -3c, EBKA-5 is called EBNA-LP (for leader protein), and LMP-2A and -2B are termed TP-1 and -2 (for terminal protein) (see also Section 1V.B). Other latent viral gene products are two highly abundant (- 105- 10' copies per cell), nonpolyadenylated, nontranslated small RNAs, termed EBER-1 and -2 (Lerner et al., 1981; Jat and Arrand, 1982), and a group of transcripts in the BumHI A region of the EBV genome (Fig. 1) whose coding capacities are uncertain (Karran ct al., 1992; Brooks et al., 1993). Three distinct patterns of EBV latent gene expression have been described in aitw in B cells, designated latency I , 11, and 111 (Table I ; Kerr et

L

GL L c;

3 LMPU

L 1 \ \ EBER-1 EBER.2

wp

BYRFl

QPiFP)

w2

3

BKRFl EBNA-1

>+ +

BLLFla BERFU BLLFlb BERF28

Y2

E LL

EBNM

LMP28

BLRF3 BERF3 BERFl BERM

E M 3 EBNAB

H

EBNA2

IklrnM)

Kilobases I 0

I

m

I 40

I

I

I

I

I

I

60

80

100

120

140

160

I (6) 172.282

FIG. 1. Genomic map of EBV. ( I ) Organization diagram of the B95-8 genome. The deletion in B95-8 in the BamHI I region relative to other EBV genotypes (e.g., AG876 and P3HR-1) is indicated. Abbreviations: TR, terminal repeat; U1-5, unique sequences 1-5; IR1-4, internal repeats 1-4. (2) Positions ofthe B95-8 EcoRI restriction fragments. (3)Positions of deletions in the P3HR- 1, Daudi, and Raji genotypes relative to B95-8 and the aforementioned deletion of B95-8 in the BamHI I region. The organization of the WZhet DNA palindrome is also shown. (4) Positions of the B95-8 BamHI restriction fragments. (5) Positions of the origin of plasmid replication (oriP), EBNA promoters (Cp, Wp, and @J (Fp))and selected open reading frames (above arrows), and their respective proteins (below arrows), as discussed in the text. The arrows indicate the directions of the respective open reading frames. Abbreviations: EBER, EBV-encoded small RNA, EBNA, EBV nuclear antigen; LMP, latent membrane protein; MA, membrane antigen. The open reading frame nomenclature is based on the BamHI viral genomic fragments; e.g., BYRFl (BamHI Y right frame one) means the first (leftmost) open reading frame starting in BamHI Y and transcribed in the rightward direction. (6) bp coordinates of the B95-8 genome expressed in kb [according to Baer et al. (1984)].

202

JAN W. GRATAMA AND INCEMAR ERNEERG

TABLE I EBV GLKEE X P R E S S IIOK~T H H LATEXCY ~L T~rt.s Cell pfienot?pe Latency type

EBER

I

I +, 2'

QJ (FP)," 1

I1

1 +,2'

QJ (@),,> 1

111

I + . 2'

W v C p , l+. 2+, 3 + , 4+,.5+, 6'

EBNA

LXIP

Burkitt's lymphoma, group I BL cells

+

+

Cell type

1 +, 2 A + , 2 B +

NPC, Hodgkin's disease

1'. 2A'. 2 B +

LCL, group 111 BLcells

E B N A m R N A transcription in latencies 1 and I1 is driven by a promoter initially thought to be situated in BninHl F (Sample el al.. 3991; ScRaefer PI nl.. 1991) but now located in BarnHl Q (Schaefer and Sprck. 1994). Therefore. this promoter is referred to as Qp ( F p ) .

ad., 1992; Rowe et al., 1992). The latency I pattern is observed in fresh

EBV+ BL tumors and in some cell lines derived from these tumors (designated group I cell lines; Rowe et al., 1987a) and consists of the expression of EBNA-1 and the transcription of EBER-1, -2, and the BamHI A mRNA fragments. In these BL tumor cells and cell lines, EBNA mRNA transcription is driven by a promoter initially thought to be situated in BarnHI F (Sample et al., 1991; Schaefer et al., 1991) but now located in BamHI Q (Schaefer and Speck, 1994). Hence, we shall refer to this promoter as Qp (Fp) in this and subsequent sections. EBV+ NPC tumor cells (Young et al., 1988; Gilligan et al., 1991) and EBV+ Reed-Sternberg cells in HD (Pallesen et al., 1991; Deacon et al., 1993) feature the latency I1 pattern, characterized by the expression of LMP-1, -2A, -2B, and @I (Fp)-driven EBNA-1 and the transcription of EBER-1, -2, and the BamHI A fragments. Finally, other (designated group 111) BL cell lines have gradually acquired, during serial passage, the latency 111 pattern, which is also the hallmark of the lymphoblastoid cell lines (LCL) derived in nitro following transformation of B lymphocytes by EBV. The latency 111 pattern is characterized by transcription of EBER-1, -2, and the BamHI A fragments and expression of the full set of nuclear (i.e., EBNA-1-6) and membrane (i.e., LMP-I, -2A, -2B) proteins. T h e EBNA mRKAs expressed in latency 111 are generated from long primary transcripts that are driven initially from the viral BamHI W (Wp) promoter and, later in the transformation process, from the BamHI C (Cp) promoter located upstream from Wp (Fig. 1). These primary transcripts are then differentially spliced to produce the individual EBNA mRNAs (Woisetschlager et (11.. 1989, 1990). T h e expression of EBV genes in in VZZKJ EBV-infected B lymphocytes

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

203

has been studied at the transcriptional level using reverse transcription polymerase chain reaction (RT-PCR) assays (Qu and Rowe, 1992; Tierney et al., 1994; Chen et al., 1995). In mononuclear cells from patients recently infected with IM, the full spectrum of investigated latent RNAs was detected, i.e., EBER-1, EBNA-1 and -2, LMP-1, and LMP-2A, with EBNA-1 mRNA spliced from the Wp, Cp, or QP (Fp) promoters (Tierney et al., 1994). In contrast, one study of mononuclear cells from long-term EBV carriers revealed no evidence of any differentially spliced EBNA transcripts but Cp-initiated mRNAs in 3 of 4 cases (Qu and Rowe, 1992), whereas in the other studies EBNA-1 mRNA was detected driven from @ (Fp) or another as yet undefined promoter (Tierney et al., 1994; Chen et al., 1995). Interestingly, EBNA-1 mRNA was detected in highdensity B cells, representing the resting subpopulation (Chen et al., 1995). EBER-1 transcripts were studied by Qu and Rowe (1992) and Tierney et al. (1994), and LMP-2A transcripts in all three studies, and were detected in all cases. The combined results of these studies indicate that, shortly following primary EBV infection, the viral gene expression pattern in B cells by and large resembles that seen in LCL, i.e., latency 111. With time, the in vivo EBV-carrying, resting B cells appear to downregulate their EBV gene expression pattern to a level similar to that of latency I. In fact, a continuum of latency states may exist for EBVinfected B cells related to their state of differentiation and activation. In this context the observations by Taylor et al. (1994a) are relevant, who described intermediate patterns of EBV gene expression in LCL with a phenotype of early or centrocytic (CDlO+) B cells, i.e., transcription of Wp in combination with QP (Fp) or Cp, and of LMP-2B but not LMP-2A mRNA and weak expression of EBNA-2 in the presence or absence of LMP- 1. Similar intermediate patterns (i.e., EBNA- 1+,2-6+,LMP- 1- and EBNA-1+,2+,3-6-,LMP-l+) have also been detected by immunoblotting of lesions of EBV+ lymphoproliferations in immunodeficient patients (Falk et al., 1990; Gratama et al., 1991). T h e latency of EBV in LCL is under tight control. Only a minor fraction (1 in 102-106 cells) in LCL cultures switch from latent to replicative infection, which is associated with viral replication and release of infectious virus (Klein and Dombos, 1973; Sugden et al., 1979). T h e switch from latent to replicative infection is *mediated by the transactivating, immediate-early Barn HI Z EBV replication activator termed ZEBRA [reviewed by Miller (1990)l. The ZEBRA protein mediates a genetic switch between the latent and lytic cycles of the virus by activating a cascade of promoters for genes that serve the production of new virus particles. T h e replicative cycle is characterized by extensive transcription of the viral genome with subsequent expression of the early

204

IAN W. GRATAMA A N D INGEMAR ERNBERG

(nonstructural) and late (structural) gene products. One of the late proteins, the major viral membrane glycoprotein (gp340/220), mediates the binding of the virion to CD2 1, the cellular EBV receptor [reviewed by Nemerow et al., (1990)l. 111. EBV Typing at the DNA Level (Genotyping)

A. INTRODUCTION

The analysis of genomic variation of infectious agents by using polymorphisms in DNA fragments generated by restriction endonuclease digestion is a powerful technique for studying the epidemiology and pathogenesis of the diseases caused by such agents. This approach was pioneered for the study of EBV by Given and Kieff (1978) and by Bornkamm et al. (1980). The discovery of significant genetic differences between EB viruses carried by two BL cell lines of West African origin, Jijoye (Pulvertaft, 1964) and AG876 (Pizzo et al., 1978), on the one hand and the prototype B95-8 virus on the other hand has greatly stimulated molecular studies of EBV infection patterns (Dambaugh et al., 1984; Adldinger et al., 1985). T h e availability and use of probes derived from the nonhomologous regions of the M-ABA (B95-8-like; Polack et al., 1984) and Jijoye (Adldinger et al., 1985) EB viruses allowed the characterization of EBV isolates as Type A (i.e., B95-8-like) or Type B (i.e., AG876-like) (Zimber et al., 1986) o r EBV-1 or EBV-2, respectively (Kieff and Liebowitz, 1990). With the advent of the PCR technique, PCR-based assays were developed, allowing direct detection of Types A and B EBV in clinical specimens. A large number of studies have subsequently addressed the presence of Type A vs B EBV in several pathological conditions, as well as in apparently healthy carriers, as discussed in the following. While the distinction between Type A and Type B allows the division of EBV isolates into t w o broad categories, discrimination between EBV isolates on the basis of deletions of DNA, the number of certain repeated sequences within the genome, and/or the presence or absence of restriction endonuclease recognition sites allows finer resolution (Bornkamm Pt al., 1980; Fischer et af., 1981; Katz et al., 1986). We shall refer to EBV isolates defined by their genomic organization as genotypes in the subsequent sections. Assessment of such patterns of variation, i.e., genotyping, has proven to be very informative for the study of wild-type EBV isolates (Hu et al., 1991; Lung et al., 1988, 1990). As an example, EBV genotypes in several cases of NPC (Hu et al., 1991), HD (Knecht et al., 1993a,b; Sandvej et a!., 1994), peripheral T-cell lymphomas (PTL), and

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

205

IN[ (Sandvej et al., 1994) share a 30-bp deletion of the BamHI N fragment. Furthermore, oropharyngeal specimens, particularly those from the lesions of hairy leukoplakia in HIV-infected individuals, may contain EElV virions with extensively deleted and rearranged defective DNA that are capable of enhancing the replication of latent parental viruses (Rabson et al., 1983; Miller et al., 1984). Finally, restriction fragment length polymorphisms (RFLP) have been shown to be useful to study EBV transmission patterns (Katz et al., 1986) and to assess whether or nolt the EBV genotypes of isolates from different sites within a single individual are similar (Bornkamm et al., 1984; Katz et al., 1988). In the next five sections, these applications of EBV genotyping will be reviewed and summarized. OF TYPES B. DIFFERENTIAL DETECTION A A N D B EBV

1. Introduction T h e ability to distinguish between Types A and B EBV genotypes stems, in the first place, from sequence divergence between the B95-8 and AG876 EBV genotypes within the open reading frame encoding EBNA-2, which is located in the BamHI YH region of the genome (Fig. 1; Dillner et al., 1985; Hennessy and Kieff, 1985; Muller-Lantzsch et al., 19185; Rymo et al., 1985). These two EBV genotypes have only 64% sequence identity through the EBNA-2 open reading frame, while the noncoding DNA outside that open reading frame, with the exception of a 105-bp deletion just downstream of the B95-8 EBNA-2 open reading frame, is 96% identical (Dambaugh et al., 1984). Second, the divergence between the B95-8 and AG876 genotypes extends to the open reading frames of EBNA-3 (BLRF3 and BERFl), EBNA-4 (BERF2a and BERF2b), and EBNA6 (BERF3 and BERF4). The B95-8 and AG876 viruses have only 90, 88, and 81% sequence identities with these open reading frames, respectively (Sample et al., 1990). In contrast, Types A and B EBV have nearly identical LMP-1 (BNLF1) genes (Sample et al., 1994). The consistency and significance of other differences between Types A and B viruses have been less well established (Arrand et al., 1989; Lin et al., 1993a). T h e results of the reviewed studies addressing the differential detection of Types A and B EBV in various clinical conditions are summarized in Table 11. 2. Studies on Cultured Cell Lines

In the initial study addressing the carriage of Types A and B EBV by cell lines derived from endemic and sporadic BL (Zimber et al., 1986),

Riu

Eiitleniic

Endemic/ Eiideinica Eiidernica Entiemic’~ Sporadic( Sporadic Sporadich Asia China China Malaysia Taiwan Southern China Northern and Central Africa Alaska U.S. mainland Southern China Western Europe Western Europe

13 I,(: I. 13I A :I,

2 (14%)

0

0

(80%) 2 (20%) 8 (500/0 x (50%) 22 (39%) 28 (50%) 15 ( I O O r r ) 0 4 3 (75%) 0 28 25 (89%) 3 ( 1 1 % ) Nasopharyngeal Carcirioma (NPC) 5 5 (100%) 0 37 32 (86%) 5 (14%)) I (3%) 29 25 (86%) 4 4(100%) 0 53 50 (94%) 2 (4%’) 1 (4%)) 25 24 (96%) 12 12(100%) 0

0

0

10

I6 50 15

131.(:1,

8X BX LCL. BX BX BX BX BX BX

0

I2

HX BX BX

SB SB SB

SB PCKT PCKr SB

k i t t i Lymphoma (131,)

6 (43%) 5 (42%)

14

6 (43%) 7 (58%)

x

0

0

:1 (5%) 0 0 0

3 (5%) 0 1(25%)

0 0 0 0 I (2%) 0 0

0

0

0 3 (10%) 0 0 0 0

3 SB 3 (100%) 0 0 0 SB 0 1 (14%) 7 5 (71%) 1 (14%) PCKr 1 (3%) 0 0 36 35 (97%) Hodgkin’s Disease (HD) in Individuals With Negative or Unknown H I V Serology LCL SB 1 1(100%) 0 0 0 BX PCRo 11 1 1 (100%) 0 0 0 BX BX BX

Zimber rt al. (1!)86) Young rt al. ( 1987) Atxtel-Hamid r / nl. (1992) (ktldschnridts rl rrl. (1992) Aitken rf nl. (1994) Zimber r/ al. (1986) Abdel-Hamid PI al. (1992) Goldschmidts rf al. (1992) Zimber rt (11. ( 1986) Hu rf al. (3991) Abdel-Hamid rt nl. 1992) Abdel-Hamid el al. 1992) Shu ct al. ( 1 992) Chen PI a/. (1992b) Abdel-Hamid et al. 1992) Abdel-Hamid et al. (1992) Abdel-Hamid et al. (1992) Choi et al. (1993) Zimber et al. ( 1 986) Gledhill et al. (1991)

Western Ecrene r-

Western Europe Western Europe Australia U.S. mainland U.S. mainland Egypt Algeria

hl

c-ALL (western Europe) NHL (Algeria) NHL (Egypt, U.S. mainland) NHL (western Europe) LMG (western Europe) AILD (western Europe) PTL (Denmark, Malaysia) PT-LPD (USA) H N T (Taiwan) PGC (Alaska) Australia Western Europe Australia OHL (Australia) United States Western Europe NHL (Australia) NHL (Western Europe)

AX BX BX BX BX BX BX BX LCL BX BX BX BX BX BX BX

! (4%) 0 24 19 (79%) 55 1 (2%) 3 (5%) 51 (93%) 0 4 2 (50%) 2 (50%) 9 0 2 (22%) 7 (78%) 16 5 (31%) 9 (56%) 2 (13%) 18 3 (17%) 3 (17%) 12 (67%) 1 0 0 1(100%) 0 15 1 (7%) 14 (93%) Other hematological malignancies in HIV seronegative individuals 1 1(100%) 0 0 SB 5 1 (20%) 0 4 (80%) PCRq SB 5 5(100%) 0 0

-ne R.-.-rt a!.

!!g~s) Sandvej et al. (1994) Lin et al. (1995) Boyle et al. (1993) Lin et al. (1993a) Lin et al. (1995) Abdel-Hamid et al. (1992) Bouzid et al. (1993)

PCR-0 PCRs PCR?, SSCP' PCRP PCR,r SSCP PCRr, SSCP' SB PCRq

PCR*, SSCP' PCRT PCRr PCRs FCRq SSCP'

3 6 12 27 24

3 3 24 24

0 (50%) (25%) (89%) (100%)

1 (33%) 3 (50%) 1 (8%) 3 (11%) 0

2 (67%) 0 4 (33%) 0 0

0 0 0

0 0 4 (33%) 0 0

Other Nonhematological Malignancies in HIV Seronegative Individuals PCRq 27 21 (78%) 2 (7%) 4(15%) 0 2 SB 2 (100%) 0 0 0 HIV-Seropositive Individuals LCL SB 3(12%) 0 26 18 (69%) 5 (19%) LCL PCR' 0 1 (3%) 33 25 (76%) 6 (18%) PBMC PCR' 0 56 15 (27%) 17 (30%) 24(43%) PCR' TW 0 30 5 (17%) 13 (43%) 12 (40%) TW PCRq 0 5 (50%) 2 (20%) 10 3 (30%) PCRr 1 LN 1(100%) 0 0 0 BX PCRt 0 0 5 (50%) 10 5 (50%) PCRr BX 3 (27%) 11 4 (36%) 0 4 (36%) BX BX

Zimber et al. (1986) Bouzid et al. (1993) Abdel-Hamid et al. (1992) Lin et al. (1995) Borisch et al. (1993a) Borisch et al. (1993b) Sandvej et al. (1994) Frank et al. (1995) Shu et al. (1992) Abdel-Hamid et al. (1992) Sculley et al. (1990) Buisson et al. (1994) Kyaw et al. (1992) Kyaw et al. (1992) Sixbey et al. (1989) Borisch et al. (1992) Boyle et al. (1991) Borisch et al. (1992) (continues)

'I'AlSLE 11

(CfJtftltfUf'ff)

EHV genotypes clet~u.tecln

'IyI'c of ;~ssay*

IYX,, 208

BM1' (1J.S.A.) (Lircliac (Atrsiralia)

'I'W PI3M(:

L!riireti Slates Uiiired Svacer I)en niark i l i a rk lkti

L.CI. LCL. 'limsil Tonsil

Uiiitetl States

TW 1'131. 'I'W

PCRs PCRu PCR'

LCL TW TW

SB PCR,, PCRP

Japan" Western Europe

Western Europe L4ustralia

Iyp H

22

(55%) 5 (50%)

0 (45%)

I:!

'Iypes A and H 0

I (10%) Bonc Marrow and Organ Transplant Recipients 2 (SS'A.) 3 (50%) lxxu fi I (17%) KRu IH 7 (39%.) 6 (33%) 5 (2H%.) l i i t ' i t ious Mononucleosis (I M) 0 SB 4 4 (loo%) 0 Kit,,

Austixliar~~

-1yp A

Tor;~l

P<:w*

10

9

#! (100%)

4 (40%)

0

0

Miscellaneous lntectious Ihorders 11 8 (73%) 1 (9%) 2 (185%) 78 33 (42%) 20 (26%) 25 (32%) 37 34 (92%)) :?I (8R) 0 Apparently 3 3 17 15 21 11

Healthy Individuals (loo%>) 0 (88%) 2 (12%) (52%) 7 (33%)

0 0 3 (14%)

Uiiknowii

Reference

0 0

(hldsctiiitidts rt nl. (1992) C)c Re rt al. (1993)

0 0

Sixlxy rt nl. ( 1989) liyaw rl d.( 1 992)

0 0

Zimber rt d.(1986) Sandvej rt d. ( 1994)

0 0 0

Sixbey rt al. (1989) Apolloni and Sculley ( 1 994) Kunimoto rt al. (1992)

0 0 0

Ziniher et al. ( 1 986) Yao et al. (1991) Apolloni et al. (1994)

Unted SL??PS Algeria

Taiwan Japan

TW TW TW TW

PCRq PCRi PCRI PCR5

34 6 26 21

17 (50%) 6(100%)

22 (85%) 20 (95%)

14 (41%) 0

3 (9%)

(4%)

3 (12%) 0

1 1

(5%)

0

0 0 0 0

Sixbey et al. (1989) B o u z i d et al. (1993) Shu et al. (1992) Kunimoto et al. (1992)

209

n Abbreviations used in this column: AILD, angioimmunoblastic lymphoma; BL, Burkitt’s lymphoma; BMT, allogeneic bone marrow transplantation; NHL, nonHodgkin’s lymphoma; c-ALL, common acute lymphoblastic leukemia; HD, Hodgkin’s disease; HNT, head and neck tumors other than NPC; LMG, lethal midline granuloma; OHL, oral hairy leukoplakia; PCG, parotid gland carcinoma; PTL, peripheral T-cell lymphoma; PT-LPD, post-transplant lymphoproliferative disease. b Abbreviations used in this column: BLCL, Burkitt’s lymphoma-derived cell line; BX, tumor biopsy or other patient-derived sample containing tumor cells: LCL, lymphoblastoid cell lines; PBL, peripheral blood leukocytes; PBMC, peripheral blood mononuclear cells; TW, throat washing. c Abbreviations used in this column: PCR, polymerase chain reaction; SB, Southern blotting; SSCP, single-strannnd conformation polymorhpism. Number of cases: in parentheses, percent of total number of cases. Central Africa, L a Reunion, and Papua New Guinea. f Kenya and Papua New Guinea. P Central Africa. Papua New Guinea. I North Africa, Asia, and Western Europe. I U.S. mainland. North and South America. 1 Women attending a clinic for sexually transmitted diseases. n Suspected of infectious mononucleosis. n Tonsilitis. According to Sample et al. (1990). 1. According to Aitken et al. (1994). 9 According to Sixbey et al. (1989). According to reference specified for that study. According to Kunimoto el al. (1992). ’ According to Lin et al. (1993a). 0

1

1

210

J A N W. C R A T A M A AND INGEMAR E R N B E R C

Types A and B EBV were detected in 21 (72%) and 6 (21%) of 29 cell lines, respectively, while two cell lines could not be classified due to genornic alterations that interfered with the hybridization of the Types A- and B-specific probes. T h e combined facts that (a) the Type B EBVcarrying cell lines were only derived from patients from central Africa, La Reunion, or Papua New Guinea and (b) all North African, Asian, and European BL cell lines, as well as LCL derived from 14 other individuals from Asia, Europe, or the United States, carried Type A EBV suggested that the occurrence of Type B virus was geographically restricted. Young et al. (1987) detected Type A in 58%and Type B in 42% of additional BL cell lines from central Africa and Papua New Guinea. i\dditional immunoblotting studies of LCL (see also Section 1V.C) derived from apparentl7 healthy carriers from central Africa and Papua New Guinea revealed Types A and B EBV in 30 and 9 of 39 cases, respectively, while these virus types were detected in 97 and only 3 of 100 LCL derived from healthy Caucasians living i n Australia or England, respectively (Young et al., 1987). In the course of these studies, it also became apparent that LCL transformed by Type B EBV grew out unusually slowly compared to Type A transformants and were difficult to expand into long-term growing cell lines (Rickinson rt al., 1987). Thus, the work of Zimber et al. (1986) and Young et al. (1987) showed that infection with Type €3 EBV could occur in any community and that the poor ability of Type B EBL' to transform B lymphocytes in uztro did not appear to interfere with its contribution to the pathogenesis of BL. The extent of systemic infection of Australian Caucasian HIV carriers with Type B EBV was studied by Sculley et al. (1990). LCL derived from spontaneous outgrowth assays of peripheral blood mononuclear cells carried Type A EBV in S9%,Type B in 19%, and both types in 12%. These results were confirmed by Buisson et at. (19941, who also found in longitudinal studies of LCL derived from ~ W OHIV+ patients that the dominant genotype (i.e., Type A or B) could change with time. The combined results o f Sculley et al. (1990) and Buisson et al. (1994) suggest that the prevalence of Type B EBV is indeed higher among HIV+ EBV carriers (25%)than among HIV- EBV carriers (3%; Young et a&.,1987). Thus, the immunodeficienc); associated with HIV infection may alter the susceptibility of these individuals to (super)infection with any of the two EBV genotypes. 3. Southern Blottiltg of Uricultured Tumor Specimens

Abdel-Harnid et al. (1992) classified the Type A or B EBV carrier status of Barti H1-digested lysates of tumor specimens of various histological types after probing Southern blots with DNA fragments specific

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

211

for Type A (Polack et al., 1984) or B (Adldinger et al., 1985). They detected Type B EBV in 20% of endemic BL tumors, while none of four sporadic BL tumors carried Type B. Fifty-two NPC tumors from Asia, Africa, or mainland United States carried Type A in all but two cases. I n contrast, all of three NPC tumors and two parotid gland carcinomas from Alaska were positive for Type B EBV. 4. PCR-Based Studies

T h e first of such studies (Sixbey et al., 1989) applied two consensus primers, which were based on the published coding sequences of EBNA-2 of the B95-8 (Baer et al., 1984) and AG876 (Dambaugh et al., 1984) viruses. In this way, an 89-bp fragment was amplified that could subsequently be classified as Type A or B using type-specific probes. Alternatively, Sample et al. (1990) employed, for the discrimination between Types A and B EBV, a set of three primers consisting of a consensus 5' primer and two 3' primers specific for Types A and B EBNA-2 or -6 coding sequences. The primers were designed so that the amplified products from Type A EBV genotypes differed from those amplified from Type B EBV genotypes (249 vs 300 bp for EBNA-2 and 153 vs 246 bp for EBNA-6, respectively). Subsequently, the specificity of the amplification was confirmed by hybridization of the DNA t0.a probe common to the two EBV types. An even more comprehensive approach was followed by Lin et al. (1993b), who classified EBV into Type A or B on the basis of three coding regions, i.e., EBNA-2, EBNA-6, and EBER. For EBNA-6, the PCR assay described by Sample et al. (1990) was used. Six of the nine type-specific point mutations in the 190-bp amplified fragment of the EBER coding region (Arrand et al., 1989) were detected on the basis of mobility shifts due to conformational changes in DNA sequences (i.e., single strand conformation polymorphism; Lin et al., 1993a). PCR assays with minor methodological modifications have been used in several studies to assess the prevalence of Types A and B EBV in most EBV-associated diseases, in relation to HIV infection and in apparently healthy EBV carriers, as summarized in the following and in Table 11. a. Hemutologzcal Tumors. Goldschmidts et al. (1992) found Type B EBV occurring in 45% of HIV+ BL cases from the United States, in 50% of endemic BL cases from Ghana, but in only 11% of sporadic BL cases from North and South America. Aitken et al. (1994) typed EBV in biopsies of endemic BL (Papua New Guinea) with previously confirmed EBV positivity as genotype A in 39% and as genotype B in 50% of cases; 5% of their biopsies contained both types, and the EBV genotypes in the remaining 5% were untypeable.

212

J A N W. GRATAMA A N D INGEMAR ERNBERG

Gledhill et al. (1991) detected only Type A in the affected EBV+ lymph node or spleen samples of unselected patients with HD. Type A EBV was also clearly predominant in a study of Danish patients with HD or I M and in Danish and Malaysian patients with peripheral T-cell lymphomas (Sandvej et al., 1994), and was the only EBV type detected in post-transplant lymphoproliferations (Frank et al., 1995). De Re et al. (1993) found Type A in all but 1 of 20 typeable HIV- HD cases, but found 50% incidence of Type B in 10 HIV+ cases. Similarly, Boyle et al. (1993) detected Type A in 78% and Type B in 22% of the HD lesions of HIV- patients, while half of the NHL lesions of Australian HIV+ patients carried Type A and the other half Type B (Boyle et al., 1991). Borisch et ul. (1992) characterized EBV present in 11 NHL lesions of western European HIV‘ patients as Type A in three cases and Type B in four cases, while four cases were untypeable. Subsequently, they detected Type A EBV in three of six cases of lethal midline granuloma, a histological subtype of PTL, and Type B in the other three cases (Borisch et al., 1993a). These seven studies had in common that dual Type A or B EBV carriage by the tumors was absent or exceptional. In contrast, Lin et al. (1993b, 1995) found 23 of 38 (61%)of HD lesions to contain Type A and 7 (18%) Type B, while 8 (21%)of these 38 lesions contained both EBV types. T h e simultaneous presence of Types A and B was also detected in one-third of the lesions of another form of PTL, i.e., angioimmunoblastic lymphadenopathy (Borisch et al., 1993b). Moreover, Bouzid et al. (1993) detected a strikingly high incidence (90%) of simultaneous carriage of Types A and B EBV in lymph node biopsies of Algerian patients with HD or non-Hodgkin’s lymphoma (NHL) that had evidence of active EBV infection as indicated by detectable titers of EA antibodies. h. hlusopliaryngeaLCarcinoma. Type A EBV was detected in most southern Chinese or Taiwanese patients with NPC or other EBV+ head and neck tumors, while Type B EBV or the coexistence of Types A and B EBV was seen only occasionall) (Chen et al., 1992b; Shu et al., 1992; Choi et ul., 1993). Pre\iousl), Hu et a / . (1991) had assessed the predominance of Type A o\er Type B EBV in a series of 37 Chinese NPC biopsies by Southern blotting. c. HIV-InfP(ted aiid HIV-UnnzfPcted EBV Carriers. Sixbey et al. (1989) addressed the carriage of Type A vs Type B EBV in throat washings of healthy adults, women attending a clinic for sexually transmitted diseases, pediatric B h l T recipients, and HIV-infected adults. The prevalence of Type B EBV in these groups was clearly higher than that in

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

213

LCL-based studies. It was detected as the only genotype in 33,9,50, and 20% of cases, respectively, while the simultaneous presence of the two genotypes was also detected in all study groups (9, 18, 33, and 50% of cases, respectively). Similarly high frequencies of Type B EBV and dual Types A and B carriage were detected in the blood of Australian HIV carriers and cardiac transplant recipients (Kyaw et al., 1992). Dual Types A and B carriage was detected in 14% of healthy Australian adults and in 32% of blood specimens from patients with suspected IM; Type B EBV was detected in about half of these samples (Apolloni and Sculley, 1994). In contrast, the exclusive carriage of Type A EBV was observed in most apparently healthy EBV carriers living in western Europe, Japan, Taiwan, and Algeria (Yao et al., 1991; Kunimoto et al., 1992; Shu et al., 1992; Bouzid et al., 1993), as well as in Japanese patients with tonsillitis (Kunimoto et al., 1992).

5. Interim Conclusions T h e various PCR-based studies agree that in tonsils or LCL derived from patients with 1M and in the throat washings of apparently healthy, long-term EBV carriers Type A is predominant, with Type B occurring in fewer carriers. Dual Type A or B carriage is quite rare (i.e., 9 of 138 (7%) individuals). This pattern is also observed in patients with various infectious diseases: dual Types A and B carriage is more common in this group (27 of 126 (21%)individuals; p = 0.002 with two-sided x 2 test). In contrast, Type B EBV clearly is more frequently detected in HIV+ individuals, carried both as a single genotype and in combination with Type A EBV. This situation is also reflected in the hematological tumors carried by HIV+ individuals; the two EBV genotypes have not been detected simultaneously in these tumors. T h e EBV genotype carriage pattern in another immunosuppressed patient group, i.e., transplant recipients, is similar to that observed in HIV+ individuals. In endemic BL, Type B is detected in -50% of cases but is rare in the sporadic cases of BL (in HIV- individuals). Type B is also exceptional in NPC, although it is more common in cases from Alaska. This situation probably reflects the general prevalence of Types B and A in the geographical areas involved (i.e., central Africa, Papua New Guinea, and Alaska vs the rest of the world, respectively). The situation with respect to HD and other EBV+ hematological tumors in HIV- patients is more complex. Here, Type A EBV is also more prevalent than Type B, but both types are simultaneously present in a minority of cases. Most of the exceptional tumors that simultaneously carried Types A and B EBV have been described in a single report from Algeria (Bouzid et al., 1993).

214

J A N W. GRATAMA A N D INGEMAR EHNBERG

C. GENOMIC HETEROGENEITY OF THE LMP-1 CODING REGION I N THE BamHI N FRAGMENT The BanzHI N fragment of the EBV genome (Fig. 3) contains the three exons of the BNLFl gene coding for LMPl (5’-3’, B95-8 coordinates 169,474 to 169,207, 169,128 to 169,042, and 168,965 to 168,163; Baer et al., 1984). T h e B95-8 LMPl protein (386 amino acids (aa)) consists of a short, cytoplasmatic N-terminal domain of 20 aa, six membrane-spanning domains, and a 200-aa cytoplasmatic C-terminal domain (Fig. 2; Liebowitz et af., 1986). T h e first 44 aa of the cytoplasma-

Frc. 2. Predicted structure of the LMP-1 protein as it would be inserted in the membrane of EBV-infected cells. Numbers in the figure refer to B95-8 amino acid positions. *: nonconservative amino acid changes detected in Alaskan and Chinese EBV isolates as detected by W.E. Miller PI nl. (1994).@: phosphorylation sites at Ser (313)and Thr (324). Interrupted lines indicate the positions of the 1 1-aa (33-bp) repeats (between aa 250 and 308) and the 10-aa (30-bp) deletion (between aa 343 and 352). Adapted from Miller et al. ( 1994).

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

215

tic domain are probably essential for cell growth transformation, while the remaining C-terminal 155 aa may provide a growth factor-like effect for B-cell transformation (Kaye et al., 1995). Hu et al. (1991) compared the DNA sequence of BNLFl and its promoter region of B95-8 (Miller et al., 1972), Raji, a cell line derived from an African BL tumor (Pulvertaft, 1965), and the nude mouse-propagated epithelial tumors CAO (derived from a Chinese NPC tumor; Cao et al., 1987) and C15 (derived from a North African NPC tumor; Busson et al., 1988). Raji was 97.5% similar to B95-8 over the BNLFl coding sequence, including the two introns (Hatfull et al., 1988), while the homology between B95-8 and C15 over that region was 99%. The CAO BNLFl gene differed from the B95-8 gene in the following ways: (a) numerous N-terminal nucleotide changes resulting in several amino acid substitutions within the first 20 aa of the encoded protein; (b) a 15-bp deletion in the C-terminal part coupled with the insertion of three 33-bp repeats, resulting in the replacement of the imperfect system of four repeats in the B95-8 and Raji genotypes by a perfect array of seven 33-bp repeats; and (c) a 30-bp deletion and several single-base mutations located 3' of the seven repeats (Hu et al., 1991). The C15 BNLF-1 gene featured one N-terminal amino acid substitution, an array of five perfect 33-bp repeats, and, 3' from the repeats, the 30-bp deletion and several single-base substitutions (Miller et al., 1994). T h e XhoI restriction site in the first exon (B95-8 coordinate 169,423 (Fig. 3)) was lost due to a point mutation in the CAO but not in the C15 tumor DNA (Hu et al., 1991). This point mutation in codon 17 consists of substitution of G in B95-8 by T in CAO, resulting in a change from Arg to Leu. Although only a single restriction site was involved, the XhoI deletion has become a useful marker to investigate the prevalence and biological significance of a novel EBV genotype in areas with a high incidence of NPC. Thus, the XhoI site was lost in 36 of 37 Chinese NPC tumors, but in only 2 of 19 African NPC tumors (Hu et al., 1991). These results were confirmed by Abdel-Hamid et al. (1992), who assessed the loss of the XhoI restriction site in 26 of 28 Chinese or Malaysian NPC tumors, but in none of 11 such tumors and in only 1 of 20 EBV+ hematological tumors from North or central Africa or the mainland United States. In contrast, all of three NPC and two parotid gland carcinoma cases from Alaska had lost the XhoI restriction site (Abdel-Hamid et al., 1992). Chen et al. (1992a) also detected the loss of this XhoI restriction site in all of 50 Taiwanese NPC tumors, but in only 1 of 6 EBV+ malignant head or neck tumors of other histologies. In a subsequent study of throat washings, they amplified the BNLFl sequence spanning

0

In (I) rY

s

8

X

v)

h 0 -

X

N

N S

X

B

MH

895-8

N

X

8

N

N

'\,

E

GSa

SM

G

S

B

X M H

CAO 0.9

0.3

1.5

1.3

1.4

0.8

0.4

1.5

0.4

0.5

Fic;. :3. Comparison of the restriction maps of the Ham141 Nhet fragments of the B95-8 and CAO genotypes. The restriction map of the €3!).5-8 HarnHI Nhet fragment and the nucleotide numbers indicated on the map are according to Baer et al. (1984).The corresponding map for CAO was established by Hu el al. (1991).The location of the EDLl promoter is indicated with a flag. Abbreviations for restriction enzymes: B, BamHI; E, EcoKI; G, BglII; H , HindIII; M, M l u I ; N, NcoI; S, SmaI; Sa, SacII;X, XhoI. Adapted from Hu et al. (1991).

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

217

the XhoI restriction site by PCR and assessed the presence or absence of that site by XhoI digestion and sequencing (Jeng et al., 1994). Typical loss of the XhoI restriction site was found in 22 of 25 (88%) PCR-positive throat washings of NPC patients, in 46 of 66 (70%)such cases of patients with other malignant head or neck tumors; in 7 of 12 (58%)such cases of patients with tonsillitis and pharyngitis, and in 10 of 25 (40%) PCRpositive throat washings of apparently healthy individuals (Jeng et al., 1994). T h e loss of the XhoI restriction site occurs independently of the Type A or B EBV genotype, but appears to be restricted geographically (Hu et al., 1991; Abdel-Hamid et al., 1992; Lin et al., 1995). T h e number of C-terminal ll-aa repeats varied between three and seven in the EBV genotypes studied (Hu et al., 1991; Chen et al., 1992a; Miller et al., 1994). PCR amplification across the LMP-1 repeats from lesions of oral hairy leukoplakia revealed multiple LMP-1 genes with different numbers of repeated sequences. Analysis of another C-terminal LMP-1 sequence that varied between unrelated EBV genotypes revealed intralesional sequence identity. Thus, the variation in the number of 11-aa repeats must have arisen by heterologous recombination during EBV replication. In a limited number of NPC, PTL, and BL cases, the number of copies was not associated with geographical origin or type of disease (Miller et al., 1994). Therefore, this polymorphism is not relevant for EBV pathogenesis. T h e 30-bp deletion and most single-base mutations in the C-terminal part of the third exon have also been detected by Chen et al. (1992a) upon sequencing the BNLFl gene of a Taiwanese NPC tumor (clone 1510) and by Knecht et al. (1993a,b) in PCR-based studies of lesions in European patients with HD. The C-terminal 30-bp deletion and single-base mutations segregated with a clinically aggressive course of HD (-10% of patients; Knecht et al., 1993b), with the transition of angioimmunoblastic lymphadenopathy into B-immunoblastic lymphoma (Knecht et al., 1995), and allowed the assessment of the persistence of the same EBV genotype in an early and a late relapse of HD (Brousset et al., 1994).The C-terminal 30-bp deletion and single-base mutations were also detected in -30% of Danish patients with HD or I M and in 11 of 18 (6 1%) Danish and in all of 9 Malaysian patients with peripheral T-cell lymphoma (Sandvej et al., 1994). Miller et al. (1994) detected the 30-bp deletion in all of 4 cases of Chinese or Malaysian NPC, 3 of 6 Mediterranean or American NPC, none of 2 Alaskan NPC, and none of 2 cases of American PTL. The C-terminal single-base mutations were also detected in the absence of the 30-bp deletion, and in some cases, 30-bp deletions were observed in the absence of single-base mutations relative to the B95-8 genotype. Also, both alter-

218

J A N W. GRATAMA A N D INGEMAR ERNBERG

ations were observed in Type A or B EBV genotypes (Miller et al., 1994; Sandvej et ul., 1994). Therefore, Sandvej et al. (1994) concluded that the 30-bp deletion and the single-base mutations occur independently and suggested that the 30-bp deletion and some of those mutations constituted hot spots for mutation in the evolution of EBV genotypes. Transfection of the BNLFl genes derived from clone 1510 and the CAO tumor into the murine epithelial cell line 3T3 (Chen et al., 1992a) and the human keratinocyte cell line Rhek-1, respectively (Hu et al., 1993), confers transformed morphology on these cell lines and renders them tumorigenic in immunodeficient mice. In addition, CAO-derived LMP- 1 has lost its immunogenicity in comparison to B95-&derived LMP-I in a murine model system (Trivedi et al., 1994). 1. Interim Conclusion

Studies of the BNLFl gene have revealed several polymorphisms with relevance for the EBV epidemiology and pathogenesis of EBV-induced malignancies. First, a point mutation leading to the loss of the XhoI restriction site in the first exon appears to be associated with an increased risk for NPC. XhoI-loss genotypes are frequently detected in NPC tumors from China, Taiwan, Malaysia, and Alaska and less frequently from Taiwanese EBV carriers without malignant disease, but are exceptional in EBV+ tumors from other parts of the world. Whether the point mutation itself has biological significance or is only a marker for a particular EBV genotype containing other, pathogenetically relevant genetic changes is unknown. This question can be addressed by studies employing the LMP-1 transfectants as described by Chen et ul. (1992a) and Hu et al. (1993). Second, a 30-bp deletion and some single-base mutations in the C-terminal part of the third exon appear to be associated with peripheral T-cell lymphoma and (aggressive forms of) HD, but they may as well represent common variants of the EBV genome in healthy carriers. Too few data are as yet available to evaluate their biological significance. T h e loss of the XhoI site occurs independently of the 30-bp deletion (Miller et al., 1994; L. F. Hu and I. Ernberg, unpublished), as do the single-base substitutions. As the loss of the XhoI site, the 30-bp deletion and the single-base substitutions are all present in two BNLF 1 genes with tumorigenic potential, as shown in immunodeficient mice by the particularly aggressive proliferation patterns of epithelial cell lines transfected with either of these genes; the crucial (combination of) sites in the BNLFl gene for EBV-induced tumorigenesis are still unknown. So far, the loss ofthe XhoI site shows the strongest association with NPC.

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

219

D. OTHERRESTRICTION FRAGMENT LENGTHPOLYMORPHISMS Lung et al. (1990) grouped EBV genotypes detected in NPC tumors into Types C and D and into prototype F and f variants. Type D EBV genotypes are characterized by the presence of a BamHI restriction site between the BamHI W1* and 11* fragments, which are lost in the prototype B95-8 genotype (Fig. 1; Raab-Traub et al., 1980), while Type C genotypes lack this site (Lung et al., 1990). The f genotype is distinguished from F by an additional BamHI site in the BamHI F fragment, which is due to a single-base mutation (Lung and Chang, 1992). Both CID and Flf polymorphisms represent only limited variability in the EBV genome in regions that are as yet not known to be active in latency or in EBV+ tumors. Although the biological relevance of these polymorphisms is therefore limited, they provide useful markers for epidemiological studies of EBV infection. T h e predominant EBV genotype in southern Chinese NPC tumors is Cf, while a minority of such tumors carry the CF genotype, which is common among apparently healthy southern Chinese EBV carriers (Lung et al., 1990). Although most apparently healthy individuals and NPC patients carry either the F or f genotype, dual infections have been detected in throat washings and blood or by comparing the EBV genotype in tumors with that in throat washings andlor blood (Lung et al., 1991, 1992). Double infection of NPC tumors has not been observed. Interestingly, the f genotype was detected in the throat washings of most of the southern Chinese NPC patients within 3 years following radiotherapy of their tumors, while the F genotype became predominant thereafter (Lung et al., 1991). This observation suggests that most NPC patients were dually infected and that the NPC tumors constituted the main reservoir of the f genotype. In contrast to southern Chinese NPC, Caucasian NPC tumors carry either DF or CF genotypes (Lung and Chang, 1992). Only DF and, less frequently, CF genotypes have been detected in LCL derived from Californian patients with IM or apparently healthy carriers (Lung et al., 1990). Chinese immigrants to the United States retain a high frequency of NPC (Buell, 1974). Virtually all first- and second-generation Chinese immigrants to California in which NPC had developed still carry the C genotype, while only half of them carry the f genotype (Lung et al., 1994). Thus, the CID and Flf polymorphisms appear to be geographically restricted, and the Cf genotype is associated with NPC in southern China.

220

J A N W. GRATAMA A N D INGEMAR ERNBERG

Abdel-Hamid et al. (1992) performed a comparative study on the A/B, C/D, F/f, and BamHI NXhofpolymorphisms of EBV in tumors of various histological types (mainly NPC and BL) from different geographical areas. Their results confirmed those of Lung et al. (1990) in that most NPC tumors from China and Malaysia carried the Cf EBV genotype, whereas the DF genotype was predominant among NPC tumors and various hematological tumors (BL, NHL, HD) from Africa (central and Northern) and the mainland United States. C or D genotypes were observed in both Types A and B viruses, while the f genotype was only detected in Type A EBV. In contrast, the C/D and F/f polymorphisms were strongly linked to the Bum HIXholpolymorphism. The most frequently observed EBV genotypes were Bum HIXholrefalned,D,F (NPC and hematological tumors from Africa and the mainland United States), Barn HI,Yho,los~,C,f (Chinese and Malaysian NPC), and Barn HIXhoI*oss,C,F (NPC and parotid gland carcinoma from China and Alaska). T h e Alaskan EBV genotype also featured loss of the BamHI site between the H and F fragments (Abdel-Hamid et al., 1992). Repeated stretches of DNA in the BLLF- 1 exon, coding for the gp340 component of the membrane antigen, a candidate EBV subunit vaccine, have been shown to vary in length between EBV genotypes, independently of their Type A o r B classification (Lees et al., 1993). Importantly, these variations d o not result in alterations in the epitopes recognized by B and T cells. T h e use of this polymorphism in epidemiological studies has not been reported.

E. DEFECTIVE EBV GENOTYPES Defective EBV was first identified in a cellular subclone of an unusual laboratory mutant of EBV, PSHR-1, which differs from its parent virus, carried by the BL-derived cell line Jijoye (Pulvertaft, 1964) by its inability to transform B lymphocytes (Miller et al., 1974). This deficiency correlates with a 6.8-kb genomic deletion in the BamHI WYH region that encodes part of EBNA-5 and EBNA-2 (Rdbson et al., 1982; Bornkamm et ul., 1982). T h e nontransforming phenotype of P3HR-1 confirms the requirement of EBNA-2 for the initiation of lymphocyte transformation (Hammerschmidt and Sugden, 1989). Moreover, P3HR- l-carrying cells also produced deleted, rearranged EBV genotypes forming selfcontained replicons able to spread from cell to cell and induce the replication of endogenous EBV upon superinfection of cells (Rabson el al., 1983; Cho et al., 1984; Miller et al., 1984, 1985a). T h e DNA of these defective EBV genotypes is arranged as a 16-kb palindrome and has the ability to induce replication maps to a 2.7-kb fragment formed by het-

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

22 1

erologous recombination between DNA sequences in BamHI W and Z, which are normally separated by >55 kb, as in B95-8 EBV (Fig. 1; Countryman and Miller, 1985; Jenson et al., 1987). This WZhet fragment contains the entire BZLFl open.reading frame encoding the ZEBRA protein, and its ability to activate EBV replication is caused by the positioning of positive regulatory elements on either side of BZLFl (Countryman et al., 1987; Rooney et al., 1988). By using PCR to amplify a 181-bp sequence spanning the abnormal junction between the Bum HI W and Z fragments in defective virions, followed by sequencing, Patton et al. (1990) detected similar but not identical sequences in 2 of 10 biopsies from oral hairy leukoplakia lesions. In a subsequent study, Gan et al. (1993) detected WZhet EBV genotypes in 6 of 16 (38%) EBV+ biopsies of oral hairy leukoplakia, in 2 of 6 (33%)EBV+ biopsies of oral cancer lesions, and in 3 of 25 (12%) EBV+ biopsies of salivary glands that were either normal or affected by Sjogren's disease. Accompanying viral replication was demonstrated by in situ hybridization and demonstration of linear EBV DNA. WZhet EBV was also detected in an epithelial thymic carcinoma that contained a minor amount of linear (replicating) EBV in addition to the major monoclonal episomal population (Patton et al., 1994). Thus, defective EBV that activates EBV replication can be present in physiological EBV infections and in EBV-associated malignancies. Sixbey et al. (1991) detected P3HR-l-like, nontransforming EBNA-2deleted EBV genotypes by PCR analysis in cell lysates from the productively infected marmoset cell line B95-8 (Miller et al., 1972), in the throat washings of 5 of 33 (15%) apparently healthy adults, and, abundantly, in biopsy specimens from 11 of 12 (92%) patients with oral hairy leukoplakia. These results, as well as the simultaneous presence of Types A and B EBV in oral hairy leukoplakia, were confirmed and extended by Walling et al. (Walling et al., 1992, 1994; Walling and Raab-Traub, 1994), who described spatially and temporally changing populations of multiple EBV genotypes in these lesions. These authors defined types as the EBV genotypes distinguished by sequence variation in the EBNA-2, -3, -4, and -6 genes (i.e., A or 1 vs B or 2), strains as EBV genotypes characterized within a type by a consistent profile of sequence variation reflected in RFLP, substrains for a given gene (e.g., coding for EBNA-2) as EBV genotypes differing from the reference strain in a few additional base substitutions in that sequence, and variants for a given gene as EBV genotypes characterized within a reference strain by different numbers of repeat elements or other genomic rearrangements in that sequence. I n a detailed PCR and sequencing analysis of the EBNA-2 coding region and its surroundings in oral hairy leukoplakia lesions, Walling et

222

JAN W. GRATAMA A N D INGEMAR ERNBERG

al. (1 994) described the presence of intact EBNA-2 genes, variations in the length of the EBNA-2 polyproline region, and a variety of deletions from the BamHI WYH region in several Types A and B genotypes (Fig. 4). T h e EBV genotypes detected in six patients (Fig. 4, A-F) were categorized by comparison with the DNA sequence and structure of reference EBV strains, i.e., B95-8 (a Type A strain derived from an American patient with IM; Baer el al., 1984), W91 (a Type A strain derived from African BL, with an EBNA-2 gene differing from that of B95-8 by 13 nucleotide substitutions and a codon insertion; Cohen et al., 1991), and AG876 (a Type B strain derived from African BL; Dambaugh et al., 1984). First, Walling et al. (1994) described three Type A strains and one Type B strain. T h e wild-type viruses from which the PCR clones LC, PL, and WL (Figs. 4A, 4B, 4D) were derived were classified as belonging to the B95-8 strain on the basis of their homology of highly conserved EBNA-2 sequences. On the same basis, clones VL and LH (Figs. 4C, 4E) were classified as belonging to the W9 1 strain and clone MT982 (Fig. 4F) to the AG876 strain. T h e MT1444 clone (Fig. 4F) featured significant sequence divergences from both B95-8 and W91 strains and was considered to represent a third Type A EBV strain. Second, classification into substrains was made on the basis of additional base substitutions in the conserved EBNA-2 sequences. Thus, patient LCs lesion carried two B95-8 substrains (i.e., LC422 vs LC1360 and LC746), while a single B95-8 substrain was detected in patient PL and a single W91 substrain in two patients: VL (clones VL947 and VL560) and LH (clone LH764). Third, a variety of internally deleted variants of EBNA-2A and EBNA-PB were detected, some of which appeared to have evolved endogenously from EBV genotypes with intact EBNA-2 genes (e.g., LC746 from LC 1360; Fig. 4A), while others appeared to be of exogenous origin (e.g., the PL, clones; Fig. 4B). In addition, endogenous recombination within EBV strains generated variants with differently sized EBNA-2 polyprnline repeats (e.g., the PL clones). I , Interm Conclusions

EBV recombination within nonrepetitive genome regions generates new genotypes during productive replication in uitro and zn viuo through the deletion, insertion, or juxtaposition of DNA sequences, which can be informative for epidemiological studies. Importantly, the reviewed cases also influence functions that are critical for the viral life cycle and, hence, are relevant for the pathogenesis of EBV-associated diseases. First, defective WZhet EB virions replicate independently, are transmissible 212 uitro, and can be detected in uwo in uncomplicated EBV infections and in EB\'-associated malignancies. Second, EBV genotypes with

223

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

C CLON E

DNA STRUCTURE mo

TYPE 6 STR AIN EBNA-ZA we1

wi

F

E CLONE

DNA STRUCTURE .no

WBI

TYPE h

TYPE 6 STRAIN

CLONE

EBNCll wm

Bow

DNA STRUCNRE PRO

,I0

A

m

STRAIPC EBNA-ZA 6W.a

EBIU-28

.Qem

FIG. 4. Sequence and structural analysis of the EBNA-2 PCR clones derived from oral hairy leukoplakia lesions in six patients (A-F). Diagrams describe the EBNA-2 type, strain, and substrain and illustrate the structure of the EBNA-2 internal deletions. The DNA sequence and structure of the HLP clones are compared with those of the appropriate reference EBV strains B95-8, W91, or AG876. Short vertical lines represent single-bp changes in sequence compared with the reference strain above it. Numbers in parentheses indicate the substrain identity of each clone based on EBNA-2 sequence variation. PRO indicates the polyproline region of EBNA-2. Deletions are illustrated by thin lines. Insertions are illustrated by small rectangles. A stop sign indicates the location of the first stop codon after a deletion in the new reading frame. From Walling et al. (1994).

EBNA-2 deletions similar to PSHR-1 have been detected in 15% of throat washings of healthy adults and in >90% of oral hairy leukoplakia lesions. Interestingly, the combined studies discussed earlier indicate that the EBNA-2 coding region is exceptionally hot for viral recombinations during replication. Thus, a rather strong pressure for selective deletion of the EBNA-2 gene must exist in the oropharynx, as it is

224

J AN W. CRATAMA A N D INCEMAR ERNBERG

difficult to reconcile the fact that EBNA-2-defective viruses survive well upon transmission and subsequent establishment of latency in new hosts. This pressure may be partly due to T-cell-mediated immune reactivity, as EBNA-2 contains peptide epitopes that are highly immunogenic upon presentation by certain class I HLA alleles (see Section V.B). F. RFLP

AS

A

TOOL TO STUDY EBV TRANSMISSION

A strategy that proved useful for the comparison of EBV isolates from epidemiologically unrelated patients consisted of digesting the EBV genome with BamHI, followed by Southern blotting and separate probing with three recombinant plasmids containing large (20-30 kb) regions of EBV DNA cloned as EcoRI fragments (Fig. 1; Katz et al., 1986, 1988). Each probe reacted with >5 different BamHI fragments, and each region contained BamHI fragments with documented RFLP in defined EBV isolates (Fischer et al., 1981; Lung et al., 1988). T h e RFLP sites were mostly encountered in regions with repetitive DNA (Fig. 1). Smaller probes reactive with individual Bum HI fragments and additional digestions with other restrictions were performed to confirm or extend results. Katz et al. (1986, 1988) demonstrated with this technique that an infant with AIDS and his mother were both infected with the same two EBV genotypes and that 5 of 7 HIV' children or pediatric liver transplant recipients could carry up to three different EBV genotypes, while 2 of 7 patients with I M carried two EBV genotypes and the remaining 5 carried a single EBV genotype. Cen et al. (1991) used the same approach, combined with a PCR assay covering the third internal repeat in the BamHI K region (coding for the EBNA-1 glycinealanine repeat), to prove that the EBV+ lymphoproliferative disease that developed in two organ transplant recipients who had received organs from a common cadaveric donor carried the same EBV genotype as the cryopreserved spleen cells of the organ donor. Of note, the proliferating, EBV-infected lymphocytes in both patients were of recipient origin. G. CONCLUSIONS

T h e analysis of genomic variation between EBV isolates has provided important epidemiological information with respect to geographical distribution and individual transmission and carriage patterns and associations between certain EBV genotypes and EBV+ malignant tumors. The most widely studied pattern of genomic variation is that of Types A and B, based on divergence of the BamHI WYH regions and extending to the BamHI L and E regions, coding for the nuclear antigens EBNA-2,

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

225

-3, -4, and -6. Types A and B EBV occur all over the world, with Type A being the most prevalent genotype. However, Type B is more prevalent in some regions (e.g., central Africa, Papua New Guinea) than in others. Numerous studies thus far have failed to show clear disease association patterns with any of the two types. T h e .fact that Type B, often accompanied by Type A EBV, is more frequently detected in immunodeficient carriers (HIV+ individuals and transplant recipients) than in immunocompetent carriers is probably due to increased exposure to exogenous virus (lifestyle and transplantation or transfusions, respectively) combined with deficient EBV-specific cellular immunity, which leads to long-term carriage of multiple EBV genotypes. Analysis of genomic polymorphisms within the BNLFl gene, coding for LMP- 1, has revealed the most promising results thus far with respect to disease-associated patterns. Transfectants containing the BNLFl genes from two Chinese NPC tumors, containing a loss of the XhoI restriction site in the first exon and a 30-bp deletion, combined with several singlebase substitutions in the C-terminal part of the third exon were tumorigenic but not immunogenic in immunodeficient mice, while the reverse was true for the LMP-1 gene of the B-cell-derived B95-8 genotype. Mutation of the XhoI restriction site appears to be associated with a risk factor for NPC, while the 30-bp deletion and single-base substitutions occur with increased frequency in patients with peripheral T-cell lymphoma o r (agressive forms of) HD. Whether one or several of these three genetic polymorphisms are directly involved in EBV-induced tuniorigenesis is unknown. T h e Bam HI, polymorphism is clearly linked to two other polymorphisms, i.e., the presence or absence of a BamHI restriction site between the BamHI W1* and 11* fragments (D vs C, respectively) and the presence or absence of an additional BamHI site in the BamHI F fragment (f and F, respectively). Thus, the predominant EBV genotype in Chinese and Malaysian NPC is BamHIx,,Loss,C,f, vs Bam HIXhoIreQined,D,F in NPC and hematological tumors from Africa and the mainland United States. However, the C/D and F/f polymorphisms appear to be geographically rather than disease restricted, similar to the Types A and B EBV. T h e extensive polymorphism of EBV genotypes is likely to be generated primarily during replication rather than in latency. The switch from the latent to replicative state of the virus is mediated by the transactivating ZEBRA protein that is encoded by the BZLFl gene. This gene is constitutively expressed in highly defective EBV virions (WZhet) that were discovered in cultures of an EBNA-2-deleted laboratory virus and subsequently in vivo in epithelial tissues. As WZhet virions can initiate the replication of endogenous viruses in latently infected cells, they may

226

J A N W. GRATAMA A N D INGEMAR ERNBERG

enhance the generation of recombinant viruses. This scenario comes into play in the lesions of oral hairv leukoplakia in HIV+ individuals that feature intense EBV replication. A wide variety of EBV genotypes have been described with respect to polyniorphisnis at the BamWYH locus (coding for EBNA-2) in such lesions. These are biologically interesting defective viruses. Finally, the extensive polymorphism of EBV genotypes allows the use of RFLP to track viral transmission patterns along natural and iatrogenic (e.g., organ transplantation) routes. IV. EBV Typing at the Protein Level (Ebnotyping)

,4.INTRODUCTION Following transformation by EBV, B lymphocytes express within 24 h a complex of EBV-encoded nuclear antigens (EBNA) that first have been defined using anticomplement immunofluorescence (Reedman and Klein, 1973). In subsequent studies, as discussed in the following, LCL were found to express six different EBNA proteins. On immunoblots, different EBV isolates can be distinguished by variations in the apparent molecular weights (MW) (hereafter referred to as size for short) of EBNA- 1, -2, -3, -4,and -6. Typing of EBV at the protein level is typically done using LCL, but it can also be performed on EBV+ tumor cells, provided that the relevant EBV-encoded protein(s) is (are) present above the detection level of the immunoblotting assay. We shall refer to EBV isolates defined by their EBNA size spectra in immunoblots as ebnotypes in the subsequent sections. IN B. EBV-ENCODED PROTEINS EBV-TRAKSFORMED CELLS

In LCL, mRNAs coding for the EBNA proteins are initially (i.e., up to 72 h postinfection) generated from the Wt, promoter (in Barn H I W) and thereafter from the C p promoter (in BarnHI C) by complex splicing from precursor RNA molecules up to 100 kb long [reviewed by Rogers ~t al. (1992)j. EBKA-I is encoded by the BKRFl exon in BamHI K, EBNA-2 by the BYRFl exon in Ba7nHI Y and H, EBNA-3 by the BLRFS and BERFl exons in BamHI L and E, respectively, EBNA-4 by the BERF2a and BERF2b exons, and EBNA-6 by the BERF3 and BERF4 exons, all in BamHI E. EBNA-5, also termed the leader protein (EBNALP) because it is expressed from the leader sequence of a bicistronic message also coding for other EBNA proteins (Wang el al., 1987), is

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

227

encoded by multiple exons derived from the BamHI W repeats and the BamHI Y fragment. The BamHI N fragment contains the BNLFl exon coding for LMP-1, while the joined termini of the episomal EBV genome serve as a template for two highly spliced messages that encode the terminal proteins LMP-2A and LMP-2B (Fig. 1).

1. EBNA-1 EBNA- 1 was first identified by immunoblotting and radioimmunoelectrophoresis in four EBV+, but not in three EBV-, cell lines by the comparison of reactivity patterns of sera from four EBV-seropositive and two EBV-seronegative individuals (Strnad et al., 1981). Transfection experiments revealed that the BamHI K fragment encoded EBNA-1 (Summers et al., 1982; Fischer et al., 1984). The B95-8 EBNA-1 protein contains a short N-terminal sequence, followed by a 20-45-kDa glycinealanine (Gly-Ala) copolymer that is flanked by basic arginine (Arg)-rich sequences, and finally a highly charged acidic C-terminal sequence (Baer ef al., 1984). A comparison of various EBV+ BL cell lines, EBV- BL cell lines infected, and LCL obtained after B-cell transformation with different EBV isolates showed that the size of EBNA-1, varying between 67 and 97 kDa, was determined by the viral isolate (Gergely et al., 1984; Sculley et al., 1984; Falk et al., 1995) and, specifically, by the length of the third internal repeat of EBV (IR3; Fig. l), i.e., the BamHI K Gly-Ala repeat (Hennessy et al., 1983; Falk et al., 1995). 2 . EBNA-2 Strnad et al. (1981) noticed an 81-kDa polypeptide, different from EBNA-1, in 2 of 4 EBV+ cell lines. A similar (i.e., 82-kDa) polypeptide with nuclear localization was described by Hennessy and Kieff (1983) in five EBV+ cell lines. Experiments using antisera directed against a bacterial fusion protein containing the BYRFl open reading frame (Hennessy and Kieff, 1985) or against synthetic peptides deduced from the sequence of that open reading frame (Dillner et al., 1985) showed that EBNA-2 is encoded by BYRF1. The B95-8 EBNA-2 protein contains a short N-terminal sequence, followed by a 26-aa long proline (Pro) polymer, farther toward the C-terminus a 12-aa Gly-Arg repeat, and finally a highly charged acidic C-terminal sequence (Baer et al., 1984). Due to the high Pro content, the apparent MW of B95-8 EBNA-2 (i.e., 82 kDa) is increased compared to the predicted MW (i.e., 55 kDa) (Perricaudet et al., 1979). The EBNA-2 proteins encoded by Type A (B95-8) and Type B (AG876) EBV share only 53% aa homology (Dambaugh et al., 1984; Adldinger et al., 1985),but the serologically recognized type-specificepitopes are in the

228

J A N W. GRATAMA A N D INGEMAR ERNBERG

relatively conserved C-terminal part (Rowe and Clarke, 1989). EBNA-PA and EBNA-2B can be distinguished by type-specific sera on immunoblots ( Y o u n g e t al., 1987; Sculley et al., 1988a). T h e size of the EBNA-2A protein varies between 85 and 97 kDa and is not determined by the sizes of the Pro and Gly-Arg repeats (Falk et al., 1995), while the EBNA-2B protein is usually represented as a doublet on immunoblots with restricted size variation around 80 kDa (Young et al., 1987). 3 . EBNA-3, -4, and -6

A serum from a patient with chronic IM (Miller et al., 1985b) reacted with a third nuclear antigen, diffeient from EBNA-1 and -2, in EBV+ but not in EBV- cell lines (Hennessy et al., 1986; Dillner et al., 1986a; Kallin et at., 1986). In subsequent studies, EBNA3 (EBNA-3a according to alternative nomenclature) was shown to be encoded by the BLRF3 and BERFl open reading frames (Hennessy et al., 1986;Joab et al., 1987; Ricksten el al., 1988). T h e size variation of EBNA-3 (i.e, 140-158 kDa) is determined by the infecting EBV isolate (Kallin et al., 1986; Falk et al., 1995). T h e B95-8 BERFl exon contains a large repeat sequence consisting of several imperfect copies of four sequences (Baer et al., 1984). T h e size of this repeat showed no significant correlation with the size of EBNA-3 in immunoblots, but the size of the BERFl coding sequence 3' of the repeat did (Falk et al., 1995). Optimization of immunoblotting conditions for high MW proteins led to the discovery of a fourth EBNA protein (Kallin et al., 1986), which was later renamed EBKA-6 (or EBNA-3c). This protein was shown to be encoded by the BERF3 and BERF4 open reading frames (Allday et al., 1988; Petti et al., 1988; Ricksten et al., 1988; Shimizu et al., 1988). Again, the size variation of EBNA-6 (i.e, 144-180 kDa) is determined by the infecting EBV isolate (Kallin et al., 1986; Falk et al., 1995). T h e B95-8 BERF4 exon contains two repeat sequences, a 10 x 15-bp and a 3 x 39bp repeat. The size of the n X 39-bp repeat shows a significant positive correlation with the size of EBNA-6 in immunoblots (Fig. 5), whereas the n X 15-bp repeat varies little, if at all, between different EBV isolates (Falk et al., 1995). EBNA-4 (or EBNA-3b) is encoded by the open reading frames BERF2a and BERF2b (Petti and Kieff, 1988; Shimizu et al., 1988; Kerdiles et al., 1990). T h e protein is -155 kDa and varies very little in size between ebnotypes (Rowe et al., 1989; Falk et al., 1995). The B95-8 BERF2b exon contains a 3 x 60-bp repeat that also shows very little variation, if any (Falk et al., 1995), while the BERF-2a exon does not contain repeat sequences. Thus, EBNA-4 has limited value for ebnotyping purposes. T h e extension of the genetic differences between Types A and B

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

229

FIG. 5. (A)Immunoblot, probed with the serum MS (M.Rowe ct ol., 1989), showing the variation of EBNA-6 in B95-8 and in 11 ebnotypes derived from healthy individuals or BMT recipients (lanes D9-Dl15). K562 is an EBV- cell line. The EBNA-6 bands are indicated with arrows. DlO8 and Jijoye M13 were not tested. (B) Southern blot after digestion with BglII (restriction sites at B95-8 coordinates 99,759 and 101,353)and probing with a DNA sequence covering the n x 39-bp repeat (ix., B95-8 coordinates 99,759 to 100,613).Jijoye M13 was not tested. (C) Southern blot of PCR products following amplification over the n X 39-bp repeat (i.e., B95-8 coordinates 100,395 to 101,119) and probing with a DNA sequence hybridizing to both Types A and B genotypes (i.e., B95-8 coordinates 100,559 to 100,618). (D) The same blot as in (C) after deprobing and reprobing with a Type B-specific DNA sequence. Numbers to the right of the figure indicate size of the proteins in kilodaltons (A) or size of the DNA fragments in kilobases (B-D). From Falk el al. (1995).

EBV from BamHI Y to BamHI E (Sample et al., 1990) is reflected in the antigenic differences between Types A and B EBV that extend to EBNA-4 and -6, but not to EBNA-3 (Rowe et al., 1989; Sculley et al.,

230

J A N W. GRATAMA AND INGEMAK ERNBERG

1989). Sera from some individuals that carry Type A EBV are reactive with EBNA-3, -4, and -6 from Type A EBV, but recognize only Type B EBNA-3. Similarly, sera derived from certain carriers of Type B EBV recognize EBNA-3 and the putative equivalents of EBNA-4 (varying between 162 and 165 kDa) and EBNA-6 (varying between 150 and 183 kDa) of Type B EBV, but react only with Type A EBNA-3 (Rowe et al., 1989). 4 . EBNA-5

Comparison of a cDNA clone from the Raji BL cell line with the B958 sequence revealed that it contained two exons of each of the Bum HI W repeats and three exons from BamHI Y , yielding a long open reading frame that, if translated, would code for a 261-aa polypeptide with a 66aa repetitive element from each Bum HI W repeat (Bodescot et al., 1984). A protein of variable size, termed EBNA-5, was identified by the use of rabbit antibodies raised to synthetic peptides corresponding to the BamHI W repeat exons (Dillner et al., 1986b). These results were confirmed by the use of a human serum that had been affinity-purified with a bacterial fusion protein expressing part of the EBNA-5 sequence (Wang et al., 1987). On immunoblots, EBNA-5-specific human antibodies detected EBNA-5 as a ladder of regularly spaced bands ranging between 20 and 130 kDa, consistent with the coding capacity of the BamHI W exons (Finke et al., 1987). However, a unique feature of EBNA-5, i.e., the expression of multiple proteins from the same BamHI W repeat region by single cells as shown by cloning of newly infected cells (Finke et al., 1987), precludes its use for ebnotyping purposes. 5 . LMP-1, -2A, and -2B

The transcription of LMPl is relatively simple compared to the complex transcription patterns of the EBNA proteins. Three closely spaced exons in BamHI N are transcribed in the leftward direction from a bidirectional promoter that is shared with LMP-2B and compose the mature 2.8-kb mRNA (Van Santen et al., 1981; Fennewald et al., 1984; Hudson et al., 1985). T h e structure of the LMPl protein is described in Section 1II.C. On immunoblots, LMP can vary in size between 57 and 66 kDa, as dictated by the infecting EBV isolate (Rowe et al., 1987b). Whether this size variation is determined by the length of the n x 33-bp repeat (Miller et al., 1994) is as yet unknown. LMP-2A and -2B are transcribed in the rightward direction; the promoter of LMP-2A is situated 3' of the LMP-1 gene. LMP-PA differs from LMP-2B in having a unique first exon encoding a 119-aa N-termi-

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

23 1

nal domain, while the remaining exons are identical for both proteins and are spliced across the terminally repeated ends of the genome (Laux et al., 1988; Sample et al., 1989). Consequently, a circularized EBV genome is required for the expression of LMP-2A and -2B. Both proteins are predicted to have 12 transmembrane domains and colocalize in membrane patches with LMP- 1 (Longnecker and Kieff, 1990). LMP-1, -2A, and -2B have not been used for typing of EBV at the protein level. LMP-1 can provide additional information to the polymorphism of EBNA-1, -2, -3, and -6 as its size remains stable in long-termcultured LCL (J. W. Gratama and I. Ernberg, unpublished). This additional information is seldom required because the use of EBNA-1, -2, and -6 was sufficient to characterize several hundred ebnotypes (Gratama et al., 1994). LMP-PA and -2B cannot be used for EBV typing as no reliable reagent for the detection of these proteins is available. Whether they vary in size therefore is unknown. C. EBNOTYPING STUDIES OF EBV CARRIER STATUS AND TRANSMISSION PATTERNS

1. EBV Carrier Status In initial experiments to characterize the EBV isolates obtained from various groups of EBV-infected individuals, LCL were established by pooling microtiter plate wells with growing cells in assays of spontaneous outgrowth (Rickinson et al., 1977) or cord blood cell transformation by EBV-containing throat washings (Ernberg et al., 1986; Sculley et al., 1987, Young et al., 1987; Gratama et al., 1990a,b). Thus, only one LCL was obtained per individual on each occasion. These studies revealed an extensive polymorphism between ebnotypes carried by unrelated individuals, particularly if EBNA-3 and -6 in addition to EBNA-1 and -2 were taken into account (Gratama et al., 1990a,b). As a rule, unrelated individuals carry different ebnotypes, the only exception being the detection of identical ebnotypes in LCL derived from the blood of one patient with Sjogren’s syndrome and from saliva directly sampled from the parotid glands of three apparently healthy individuals (Oosterveer et al., 1993). In contrast, identical ebnotypes can regularly be detected in LCL derived from family members (Fig. 6; Gratama et al., 1990b). The vast majority of LCL carry Type A EBV. Type B EBV was only carried by LCL derived from 3 of 100 (Young et al., 1987) and 1 of 37 (3%)apparently healthy Caucasians (Gratama et al., 1990a,b),but by 9 of 39 (23%) LCL derived from healthy inhabitants of Kenya and Papua New Guinea (Young et al., 1987).

232

J A N W. GRATAMA A N D INGEMAR ERNBERG

Family3

Family 2

I

28

29

33

35

28 r

24

23

18 d

FIG. 6. Pedigrees of two families and immunoblot of established LCL from their memhers, probed with the polyspecific Type .4-specific serum PG. Males and females are indicated by squares and circles: their ages at the time of initial investigation are indicated below these symbols (r, bone marrow recipient; d, bone marrow donor). Letters above the lanes correspond with those in the pedigrees. j1 and 52: LCL derived from blood obtained at day 9 prior to BMT and day 1820 post-BMT. Mother B and sibling M were EBVseronegative. Siblings D and E were not studied. Numbers to the right of the immunoblot indicate MW in kilodaltons. The Ramos BL cell line was used as EBV- control. EBNA (E)-I, -2, -3, and -6 were assigned to their corresponding bands after probing with monospecific sera. EBNA-4 was assigned by exclusion as no EBNA-4-monospecific serum was dvailable. Letters below the lines denote different ebnotypes. From Gratama el al. (1990b).

The coexistence of multiple ebnotypes in individual carriers, already suggested by Sculley et aE. (1987) as a possible explanation for the occurrence of multiple EBNA-1 and -2 bands in immunoblots of some LCL, was addressed in several studies summarized in Table 111. In contrast to

TYPES A

AND

TABLE 111 B EBNOTYPES I N VARIOUS CLINICAL CONDITIONS Ebnotype

Total

Type A

Type B

Types A and B

Reference

24 108

19 (79%) 102 (94%)

5 (21%) 6 (6%)

0 0

Infectious mononucleosis

7

7 (100%)

0

0

Yao et al. (1991) Gratama et al. (1994) Gratama et al. ( 1994)

BMT recipients prior to BMT BMT recipients post-BMTa Cardiac transplant recipientsb HIV-infected individuals

9

Patient population Apparently healthy individuals

a

b

23 20

8 (89%) 19 (83%) 18 (90%)

1 (11%) 3 (13%)

18 1

12 (67%) 0

3 (17%) 0

Median time post-BMT, 16 months (range 5-113 months). Median time posttransplantation, 17 months (range 2-42 months).

1 (5%)

0 1 1

(4%) (5%)

3 (17%) 1(100%)

Gratama et al. (1994) Gratama et al. (1994) Gratama et al. (1994) Sculley et al. (1990) Gratama et al. f 1994)

234

J A N W. GRATAMA A N D INGEMAR ERNBERG

the initial studies, microtiter plate wells with growing cells were expanded individually rather than being expanded as pooled cell suspensions. The simultaneous carriage of Types A and B EBV was restricted to immunosuppressed individuals (i.e., bone marrow of cardiac transplant recipients and HIV-infected individuals). Importantly, imniunoblotting allows the detection of multiple ebnotypes within Types A and B EBV by virtue of the discriminative power of the size polymorphisms of EBNA-1, -2, and -6 (Sculley et ul., 1990; Yao ut al., 1991; Gratama et al., 1994). In such studies, cases from which two or more LCL could be established from peripheral blood or throat wash cultures are considered informative. In a limited study, Sculley et al. (1990) did not detect multiple ebnotypes within Types A and B EBV in HIV-infected individuals. However, in a much larger study the frequency of multiple ebnotypes was even detected in healthy individuals, patients with IM, or those awaiting a BMT (blood, 1575,and throat washings, 24%,of individuals), whereas it was even more frequently observed in immunosuppressed individuals such as recipients of bone marrow or cardiac allografts (blood, 35%, and throat washings, 6992, o f individuals) (Gratama et al., 1994). Three patterns of simultaneous carriage of multiple ebnotypes were detected. T h e first pattern consisted of minority ebnotypes differing from the majority ebnotype in only a single EBNA protein (mostly EBKA-1) and was observed in all study groups. This pattern was also observed by Yao et al. (1991) in 9% of healthy EBV carriers. A second, less frequent pattern featured minority ebnotypes differing from the majority ebnotype by 2 of the 3 EBNA proteins studied (EBNA-1 and -6 in most cases) and was observed in healthy carriers and (candidate) transplant recipients. The third pattern was characterized by the simultaneous presence of ebnotypes differing in EBNA-1, -2, and 6 and was only observed in immunosuppressed individuals (i.e., transplant recipients and HIV-infected). The extensive size polymorphism of EBNA-1, -2, and -6 between unrelated individuals suggests that the first two patterns arise from heterologous recombinations during viral replication, while the latter pattern would reflect the carrier status after multiple infections with exogenous viruses (Gratama et al., 1994). Additional support for heterologous recombination of endogenous EBV was obtained by DNA analysis of seven LCL, derived from a single throat wash culture, that carried ebnotypes differing from each other in EBNA-1 only. Distant EBV sequences around the LMP-1 promoter in BamHI N were >99% homologous in the seven LCL, whereas these sequences showed only 90-95% homology in EBV genotypes derived from unrelated individuals (Friis et al., 1995).

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

235

2. EBV Transmission Patterns T h e combined use of ebnotyping and EBV serology has been particularly informative to investigate the origin and kinetics of EBV infection in recipients of bone marrow (Gratama et al., 1988, 1990a, 1992, 1994) or kidney allografts (Van Gelder et al., 1994). These studies led to the following conclusions. (a) EBV is frequently transferred with the allograft. This transfer is illustrated by the case of an EBV-seronegative recipient who received a T-cell-depleted marrow graft from an EBV-seropositive, HLA-identical, unrelated donor following pretreatment with cytotoxic drugs and total body irradiation as therapy for his myelodysplastic syndrome. At 2 months post-BMT, severe pharyngitis with high oropharyngeal EB virus titers developed, which gradually subsided upon administration of antiviral therapy using acyclovir. The donor ebnotype was exclusively detected in L,CL established from throat wash and peripheral blood cultures during the subsequent 30-month followup period BMT and differed from the ebnotypes of his parents and sister (Gratama et al., 1994). In an EBVseronegative kidney transplant recipient, an almost fatal EBV+ lymphoproliferation that developed at 6 months posttransplantation was traced back to the kidney donor. The lymphoproliferation regressed upon reduction of the immunosuppressive therapy that had been installed to reduce antigraft alloreactivity (Van Gelder et al., 1994). (b) EBV can be eradicated from BMT recipients. This observation has been made in two recipients of HLA-identical sibling donor bone marrow not depleted of T cells, following cytoreductive treatment for acute leukemia in remission. Following BMT, the first patient became seronegative for 4 years and was subsequently reinfected with EBV, expressing an ebnotype identical to that of her husband (Fig. 6). In the second patient the pretransplant ebnotype was persistently replaced by that of his marrow donor from day 252 onward. Both patients had clinical evidence of donor antihost alloreactivity (graft vs host disease) and are for 12 and 13 years in complete remission of their leukemia, respectively (Gratama et al., 1988). (c) EBV can persist following allogeneic BMT. This observation has been made in three recipients of T-cell-depleted, HLA-identical sibling donor bone marrow, following cytoreductive treatment for chronic myelogenous leukemia in the chronic phase. None of the patients had clinical evidence of graft vs host disease. The first patient’s marrow graft failed to engraft durably and an EBV+ lymphoproliferation developed upon therapy with anti-T-cell antibodies to eliminate host vs graft al-

236

J A N W. GRATAMA A N D INGEMAR ERNBERG

loreactivity. T h e proliferating B lymphoblasts were of recipient origin and carried the recipient's pre-BMT ebnotype (Gratama et al., 1990a). Residual hematopoietic cells were also present in the second patient, as evidenced by the cytogenetic relapse of his leukemia. T h e third patient received marrow from her EBV-seronegative brother. Oropharyngeal EBV cultures prior to and during the first month post-BMT were positive and yieided the pretransplant ebnotype. She is still in remission 4 years post-BMT (Gratama et al., 1994). T h e persistence of her pretransplant ebnotype may be explained by the infection of donor B cells by EBV replicating in the oropharynx during the peritransplant period. Alternatively, surviving residual recipient B cells (Gerhartz et al., 1988) may have re'leased EB virions that have subsequently infected epithelial cells and B cells originating from the marrow graft. From these data, two factors appear to determine eradication vs persistence of recipient EBV following allogeneic BMT. First is the presence or absence of continuous viral replication in the oropharynx. High titers of transforming EBV have been detected in the throat washings of 1025% of EBV-seropositive patients during the peritransplantation period (Gratama et al., 1992 and unpublished). Second is the presence or absence of donor antihost alloreactivity that may eliminate the recipient B cells surviving the cytoreductive therapy prior to BMT. D. CONCLUSIONS T h e methods to study EBV heterogeneity at the protein level have evolved simultaneously with the DNA techniques applied for the same purpose (Section 111). Genetically, the EBNA polymorphism is by and large determined by the size variability of repeat sequences in the respective open reading frames. These ebnotyping studies have been restricted to latently infected, i.e., transformed, cells that expressed EBNA proteins in sufficient quantities to be detectable by immunoblotting. An obvious restriction of such studies is that they provide information only on EB virions with transforming potential. Moreover, these studies are skewed against virions with low transforming efficiency such as Type B vs Type A EBV. This bias is clearly illustrated by the higher frequencies of Type B EBV in PCR-based studies (Table 11) than in LCL-based studies (Table IV) of apparently healthy individuals. However, the EBNA polymorphism can be studied at the DNA level by PCR, as evidenced by EBNA-6 PCR over the n x 39-bp repeat (Figs. 5C, 5D; Falk et al., 1995). Nevertheless, ebnotyping studies using immunoblotting have yielded several important pieces of information. First, ebnotyping revealed extensive polymorphism among field iso-

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

237

lates of EBV from unrelated individuals. That characteristic permitted the tracking of EBV infectivity patterns in natural situations and in the transplantation setting. Second, cloning experiments of LCL derived from peripheral blood and, particularly, throat wash cultures revealed significant polymorphism of ebnotypes within individual EBV carriers (Gratama et al., 1994). Thus, the genetic heterogeneity of EB virions within individuals with intense viral replication (i.e., oral hairy leukoplakia) at the BamHI WYH region (Walling et al., 1994) has its equivalent in EBNA protein size that extends from immunosuppressed individuals to apparently healthy carriers. V. Differential Recognition of EBV Genotypes by the Immune System

A. EBV GENOTYPE-SPECIFIC ANTIBODY RESPONSES T h e known serological recognition of different EBV genotypes is restricted to antibodies that distinguish between Types A and B EBV. As outlined in Section IV.B, serologically recognized type-specific epitopes have been identified in the otherwise relatively conserved C-terminal part of EBNA-2, while the polyproline repeat does not carry such epitopes (Rowe and Clarke, 1989). The existence of type-common serological epitopes on EBNA-2 is expected on the basis of 53% aa homology between the B95.8 (Type A) and AG876 (Type B) genotypes (Dambaugh et al., 1984; Adldinger et al., 1985) and proven by the generation of a monoclonal antibody recognizing both Types A and B EBNA-2 (Young ct al., 1989). In addition, Type A-specific antibodies d o not react with Type B EBNA-4 and -6 and vice versa, whereas they are cross-reactive with respect to EBNA-3 (Rowe et al., 1989; Sculley et al., 1989). T h e occurrence of antibodies specific for Type A or B EBV has been studied in sera from several groups of EBV carriers (Table IV) by immunofluorescence of EBNA-2A- or -2B-transfected cell lines (Buisson et al., 1994), immunoblotting (Sculley et al., 1988b; Yao et al., 1991; Oosterveer et al., 1993; Buisson et al., 1994), or enzyme-linked immunosorbent assays (Geertsen et al., 1994). The reactivity with EBNA-2A of most sera of apparently healthy EBV carriers and patients with primary Sjogren’s syndrome exceeds the reactivity with EBNA-2B, while these reactivities were similar in 45% of patients with IM. Only small proportions of sera react more strongly with EBNA-2B than with EBNA-2A; this pattern is seen more frequently in HIV-infected patients than in apparently healthy EBV carriers. Classification of the EBV types in the blood or throat washings of some carriers (Table IV) into Type A or B by ge-

H I V-Sertiposit ive lndividuals AIDS'/ No AIDS" Wcs t ern

Eli t-opc

IB I8 I F and I W

33 HX

43

123 146

58 (26%:)

63 (43%)

6 (14%') 20 (16%)) 9 0 (2 1%)

4 (9%)

15 (12%) 15 (10%)

Sclllley f't fAl. 1YNHb) Sculley ct al. 198%) Buisson Pt d [ 1994)

O t h e r Patient (;roups i t1v

~

hr"

M. Sjijgrenl' 1 Mo Aiidralia Wesrern Europe Western Europe Western Europe

IU

X:3

I I3

18

E.1 .lSA

5li

IB IB

ti5

IF and 1B ELISA

76 49

I04

4 (5%,) 4 (22%:) 11-1 (Y4%) I (2%) Appdrrtitly H ~ i h h EHV y I:arricrs (i0 3 (5%) 12 (16%) 4 (5%) 60 (79%) 13 (27%) 34 (69%) I (2%) 7 (7%) 79 (76%) 5 (5%)

72

3 (17%') I1 (2W)

1 I (61%)

25 (45%)

Sculley c't I d . I o w ) Oosterveer r 91. (l!)Y3) (;rrrtseii PI al ( I O!)4)

2 (3Si) 0 1 (2%) 13 (13%)

Sc iilley et al ( I 988L) Yao el al. ( 1 99 I ) Buisson et al. (1994) Geertsen et al. (1994)

7 (8%) 0

Alrbrwiirtions used in this column H I \ I - ' hr, IlIV-scrwcgativc irrriividualu at high risk of contracii~~g I Ilk' (i-c.,lwrru)suxuals,iv drug users); IM. infcc.litrrls mononucleosis; M. Sjogren, primary Sjogren's syndrome. Abbreviations used in this column: ELISA, enzyme-linked immunosorbenr assay; IB, immunoblotting; IF, immunofiuorescence. Number of cases; in parentheses, percent of total number of cases. d Australia. Western Europe. I'

MOLECULAR EPIDEMIOLOGY OF EBV INFEC r I O N

239

notyping o r ebnotyping matched the serotyping results. Overall, serotyping of EBV carriers into Type A or B (Table IV) yields an epidemiological pattern similar to that observed in LCL-based studies (Table 11). This similarity suggests that the Type A vs Type B antibody profile is a reflection of the carriage status of transforming EB viruses. B. EBV GENOTYPE-SPECIFIC CELLULAR IMMUNE RESPONSES T h e regression of outgrowth of in vitro EBV-infected B-cell cultures mediated by lymphocytes from EBV-seropositive but not sero-negative donors was the first evidence of EBV-specific T-cell immunity (Moss et al., 1978). These lymphocytes were characterized as memory cytotoxic T cells (CTL) that recognized an operationally defined, EBV-encoded, lymphocyte-determined membrane antigen (LYDMA) in the context of class I and, less frequently, class I1 HLA antigens (Wallace ef aL, 1982; Misko et al., 1984). T h e isolation of CTL clones able to distinguish between EBNA-2A and EBNA-2B indicated that the EBNA proteins were important candidates for LYDMA (Moss et al., 1988). Indeed, the occurrence of CTL responses specific for Type A or B EBV-encoded antigens corresponded with the viral carrier status of cardiac transplant recipients and apparently healthy controls, as documented by genotyping for Types A and B EBV (Kyaw-Tanner et al., 1994). T h e availability and use of a series of recombinant vaccinia viruses encoding all EBNA and LMP proteins have allowed the identification of target antigens for EBV-specific CTL restricted through a variety of HLA alleles (Murray et al., 1992; Gavioli et al., 1992; Khanna et al., 1992). As summarized in Table V, these experiments revealed that various HLArestricted, EBV-specific CTL reactivities are usually present within individual EBV carriers. T h e majority of these reactivities are indeed directed against EBNA proteins, i.e., EBNA-3, -4, and 6, while EBNA-1specific CTL responses were consistently absent (Table V). Analysis of the fine specificity of HLA-restricted, EBV-specific CTL clones at the peptide level revealed multiple CTL epitopes on some EBNA proteins (Gavioli et al., 1993; Burrows et al., 1994). Some of these epitopes varied between EBV genotypes (e.g., the HLA-B8-restricted EBNA-3 epitope TETAQAWNAGFLRGRAYGIDLLRTE of B95.8 EBV featured an I+L amino acid substitution in some Type A genotypes and an I+Q substitution in some Type B genotypes), while others were conserved (e.g., the HLAB44-restricted EBNA-6 epitope EENLLDFVRFMGVMSSCNNP). T h e single amino acid substitutions in the EBNA-3 TETA epitope had profound effects on their recognition by CTL: the I+L substitution resulted

240

JAN W. GRATAMA A N D INGEMAR ERNBERG

TABLE V SIJMMARY OF EBV ANTIGEN/CLASS I HLA COMBINATIONS YIELDING CTL TARGET EPITOPEV

-, no plyclonal or clonal C T L responses detected in any of 31 donors tested; +, weak CTL responses in some 01 the donors; + + , strong responses in polyclonal CTL cultures and rare clones of some donors; + + +, sti-ong plyclonal and clonal CTL responses in all donors. From Masucci ef al. (1993). '1

in a 15-fold increase in efficiency as the T-cell epitope, while the I+Q substitution abolished that activity (Apolloni et al., 1992). On the contrary, single amino acid substitutions in the HLA-A2. 1-restricted, LMP-'LA-encoded epitope CLGGLLTMV had no effect on CTL recognition. Here, a C - 4 substitution was geographically restricted to EBV isolates from Southeast Asia and Papua New Guinea (Lee et al., 1993). In HLA-A1 1 EBV carriers, CTL responses are frequently dominated by epitopes encoded by Type A EBNA-4 (Gavioli et al., 1992). Type '4 EBNA-4 contains several HLA-A1 1-restricted CTL epitopes, the most immunodominant of which is IVTDFSVIK (B95.8 aa residues 416-424; Gavioli et al., 1993). Substitutions at residue 424 (i.e., K+T, K-R, or K-N) were detected in 23 EBV isolates from Papua New Guinea and southeast China, while 10 other Chinese isolates had a V+L substitution at residue 417. These substitutions interfered with binding of the nonamer to HLA-A1 1 and, consequently, presentation of these epitopes at the cell surface. About half of these isolates also had a mutation in an anchor residue of a second EBNA-4 epitope. Such HLA-A11 CTL epitope-loss EBV genotypes were exceptional among 30 isolates from Europe and the United States and absent among African isolates. Interestingly, the occurrence of epitope-loss genotypes correlated strongly with geographical differences in HLA-A 1 1 antigen frequency: 50-60% among inhabitants of Papua New Guinea and Southeast Asia vs 0- 10% among Africans and Caucasian Europeans and Americans. +

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

24 1

Thus, HLA-restricted CTL responses appear to have driven EBNA-4 epitope variation, resulting in the selection of EBV genotypes lacking immunodominant HLA-A1 l-restricted epitopes in highly HLA-A1 1 populations (De Campos-Lima et al., 1993, 1994; Lee et al., 1995) may only be obtained under exceptional circumstances. +

VI. Final Conclusions: Implications for the Biology of EBV Infection Molecular epidemiological studies of EBV have made several important contributions to o u r understanding of the relationship between the virus and its host in health and disease, which are summarized kaleidoscopically in Fig. 7. A. EPIDEMIOLOGY: SORTING EBV GENOTYPES A N D EBNOTYPES IN ORDEROF THEIR BIOLOGICAL SIGNIFICANCE. T h e presence of EBV-like, B-lymphotrophic herpes viruses in all Old World monkey species studied, showing at least 40% homology with EBV at the DNA level [reviewed by Dillner and Kallin (1988); Li et al., 19941, indicates that these viruses have coevolved with their hosts over millions of years. Genetic analyses of human EB viruses occurring in humans have revealed two groups, designated Types A (or 1) and B (or 21, that were only 64-9051 homologous in sequences coding for proteins rich in CTL epitopes, while the flanking introns were 96% homologous, suggesting a common ancestor. A number of additional polymorphisms have developed independently in Types A and B viruses. Several of them are in the BamHI N fragment, which has been under intense scrutiny because it encodes the LMP-1 protein with transforming capability. These polymorphisms include base-pair substitutions relative to the reference EBV genotype B95-8 in the N-terminal part of LMP-1, deletion of the XhoI restriction site of which has been studied most extensively, and similar C-Terminal substitutions and a deletion of a 30bp stretch. Additional polymorphisms elsewhere in the EBV genome, such as C/D and Flf, are created by single-base mutations in noncoding sequences. These polymorphisms, including the major Type A/B polymorphism, show a geographical rather than disease-related distribution, with the possible exception of the XhoI-loss polymorphism (see the following). Many EBV proteins (e.g., most EBNA proteins, LMP-1 and the gp340 membrane antigen) contain repeat sequences that vary in length be-

242

J A N W. GRATAMA A N D INGEMAR ERNBERG

Healthy Individuals

:,

M. Hodkin, Peripheral T-cell Lymphoma '.:

FIG. 7. Kaleidoscopic view of the interaction between EBV and its hosts in health and disease. In the Healthy Individuals part, the transitions between the three forms of latency and replicative infection, as well as T-cell immune surveillance and neutralizing antibodies, are detailed. In the other (disease) parts, only the main deviations from the situation in healthy individuals are depicted. Interrupted contours of infected cells and hatches in epithelial cells denote replicative infection. Closed hexagons are complete EB virions, and open hexagons are defective virions (e.g., WZhet). Abbreviations: Y, neutralizing antibody; I, 11, and 111, latency forms I, 11, and 111; I*, latency I with cytogenetic accident [i.e., t(8:14). t(2:8), or t(8:22)]; CTL, HLA-restricted, EBV-specific cytotoxic T lymphocyte; PTL, peripheral T-cell lymphoma.

tween different isolates. T h e variations of the EBNA-1, -2, -3, and -6 genes were closely correlated with the sizes of the respective proteins in immunoblots, and the combined size profile of multiple EBNA proteins (ebnotype) constituted a useful marker for studies of EBV transmission and carrier status. These polymorphisms are caused by heterologous recombinations of linear viral DNA during replication, usually in repetiri\e coding sequences. Hence, multiple ebnotypes could be detected in the saliva and, to a lesser extent, in the blood of EBV carriers at a given time point, but only the dominant ebnotype was consistently detected in

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

243

longitudinal studies (Yao et al., 1991). That result indicates that such variant viruses, as a rule, represent only a minority of the viral burden in the long-term reservoir and that their biological significance is restricted to salivary transmission. The evolution of such variants over prolonged periods of time may constitute the basis of the extensive size polymorphism of most EBNA proteins in the general population. T h e nomenclature proposed by Walling et al. (Walling et al., 1992, 1994; Walling and Raab-Traub, 1994) is useful for the classification of EBV genotypes and ebnotypes and ranks these polymorphisms on the basis of their biological significance. Thus, the first-line classification is into the already defined Types A (or 1) and B (or 2) (Zimber et al., 1986; Sample et al., 1990). Second, strains are EBV genotypes characterized within a type by a consistent profile in sequence variation reflected in C,f strain in Chinese and Malaysian NPC). RFLP (e.g., the BarnHIxhrlo~~, Third, within strains, substrains for a given gene are defined as EBV genotypes differing from the reference strain in a few additional base substitutions in that sequence (e.g., the B95-8 and W91 EBNA-2A substrains shown in Fig. 4). Fourth, within (sub)strains, variants for a given gene are defined as EBV genotypes (or ebnotypes) characterized within a reference strain by different numbers of repeat elements or other genomic rearrangements in that sequence (Walling et al., 1994; Walling and Raab-Traub, 1994; Miller et a/., 1994; Falk et al., 1995). Importantly, the performance of such a comprehensive analysis at the DNA level avoids selection against poorly or nontransforming viruses that may be biologically revelant (see the following). B. EPIDEMIOLOGY: VIRALEVOLUTION UNDER IMMUNOLOGICAL PRESSURE? T h e detection of CTL epitope-loss EBV genotypes in populations with limited HLA polymorphism suggests a role for HLA-restricted, EBV-specific T-cell immunity in the genetic evolution of EBV. Case in point are the base substitutions in the BERF-2B exon, resulting in amino acid replacements that interfered with the presentation of immunodominant EBNA-4 epitopes by HLA-A1 1. The in vitro observation that no HLA-A1 l-restricted CTL responses can be generated by such carriers against their resident EBV supports the hypothesis that such EBV genotypes have a selective advantage to persist in their hosts (De Campos-Lima et al., 1994). A similar mechanism may be operative in the base substitutions in the BNLF-1 exon and the resulting amino acid replacements in the N-terminal part of LMP-1, which are also highly prevalent among Chinese EBV genotypes. An important issue is wheth-

244

J A N W. CRATAMA AND INGEMAR ERNBERG

er major differences between EBV genotypes (e.g., Type A vs Type B EBNA-2) have arisen under T-cell-mediated pressure. In this context it is relevant that EBNA-2 has a pivotal role in the regulation of EBV gene expression and B-cell activation [reviewed by Rogers et al. (1992); Kempkes et al., 19951 and that EBNA-2 deletion mutants frequently emerge (Walling et al., 1994). These combined observations suggest strong immunological pressure on this protein.

C . TRANSMISSION: SUPPORTFOR A N D CHALLENGE OF EXISTING CONCEPTS Prior to the availability and use of molecular techniques to characterize EBV isolates, the concept that EBV could be transmitted by salivary contact or, incidentally, by blood transfusion or organ transplantation relied on serological, clinical, and epidemiological studies. At that time, oropharyngeal epithelial cells, featuring chronic low-grade EBV replication in apparently healthy carriers, were considered to be the main reservoir (Rickinson et al., 1985). Subsequently, genotyping and ebnotyping studies provided direct evidence for the transmission of EBV by salivary contact and allogeneic bone marrow and organ transplantation. A combined serological and ebnotyping study of two BMT recipients revealed that successful BMT could eradicate the recipient’s resident strain and either replace it with the bone marrow donor’s strain or result in EBV seronegativity with susceptibility to salivary reinfection (Gratama et al., 1988). These results, combined with the observations that treatment with the viral thymidine kinase inhibitor acyclovir interfered with oropharyngeal EBV replication without reducing the number of circulating EBV-carrying hematopoietic cells (Yao et al., 1989), challenged the predominant role of the epithelial cell compartment in virus persistence and put emphasis on the hematopoietic Compartment as the long-term EBV reservoir (Klein, 1989; Masucci and Ernberg, 1994).

NEWTRICKS FROM D. PATHOGENESIS:

AN

OLDDoc

This allusion to the title of a recent editorial by Liebowitz (1995), commenting on the first descriptions of EBV in smooth-muscle tumors, illustrates the ever-increasing range of neoplasms that are associated with this virus. T h e quest for pathogenic EBV genotypes has thus far revealed two interesting groups: defective viruses and genotypes with base substitutions or deletions in the BanzHI N fragment. Defective EBV genotypes (WZhet) probably arise during virus replication and have been detected in the oropharynx of some normal carriers and in one thymic epithelial tumor (Patton et al., 1994). WZhet vir-

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

245

ions initiate the replication of latent virus by virtue of their constitutive expression of the transactivating, immediate-early ZEBRA protein. This characteristic contributes to the pathology of oral hairy leukoplakia lesions, which are the foci of intense EBV replication on the lateral sides of the tongue in HIV-infected patients. Whether the absence of neutralizing antibodies in severely immunosuppressed patients allows for the development of such foci, and for the dissemination of replicating EBV to other nonhematopoietic tissues such as smooth-muscle cells, is unknown. Neutralizing antibodies are maintained in healthy EBV carriers throughout life (Rocchi et al., 1973) and may constitute an initial and probably efficient barrier against (super)infection with EB virions. T h e cytoplasmatic domains of LMP- 1 contain functions essential for cell growth transformation, with a growth factor-like effect for transformed B cells (i.e., the N-terminal44 aa and C-terminal200 aa, respectively; Kaye et al., 1995). Thus, alterations in the latter part of LMP-1, such as the 30-bp deletion in the BNLF-3 exon and C-terminal single amino acid substitutions, could affect the function of the protein. The unphosphorylated form of LMP-1 has been shown to be biologically active (Moorthy and Thorley-Lawson, 1993). Thus, inhibition of phosphorylation by substituting amino acid residues that normally can be phosphorylated could increase the ability of LMP- 1 to alter cellular growth properties due to its being in an constantly unphosphorylated active form. The detection of such deletions and substitutions in EBV+ tumor featuring the latency I1 program of gene expression (NPC, HD, and PTL) therefore is relevant. As these genetic alterations have also been detected in EBV genotypes of apparently healthy carriers, their precise role in the pathogenesis of NPC, HD, and PTL is as yet unknown. E. PATHOGENESIS: LACKOF EFFECTIVE IMMUNE SURVEILLANCE T h e role of T-cell-mediated immunity in the in uzuo elimination of EBV-transformed B cells, featuring the latency I11 program of gene expression, is illustrated by the acute rejection of such cells in IM and by the development of EBV+ lymphoproliferative disease in severely immunodeficient patients and the cure of such lymphoproliferations in BMT recipients by leukocyte infusions of their EBV-seropositive marrow donors (Papadopoulos et al., 1994). Studies of apparently healthy EBV carriers have shown that this immunity is mainly class I HLArestricted and directed against EBV proteins expressed in the latency 111 program. HLA-restricted, EBV-specific CTL are probably continu-

246

J A N W. GRATAMA AND INCEMAR ERNBERG

ously engaged in the elimination of EBV-transformed B cells. The possibility that such cells may be continuously generated and provide a chronic antigenic stimulus is suggested by the relatively high frequency of EBV-specific CTL precursors in apparently healthy carriers (Rickinson et al., 1981; Bourgeault et al., 1991). The failure to mount HLA-restricted, EBN A-l-specific CTL responses by all healthy, immunocompetent EBV carriers studied is probably pivotal for the long-term persistence of EBV in small resting B cells that only express EBNA-1 (Chen et al., 1995). In contrast with LMP-1, EBNA-1 also is not recognized by murine T cells, suggesting a defect in processing o r transport (Trivedi et al., 1994). Indeed, Levitskaya et al. (1995) recently demonstrated that the EBNA- 1 Gly-Ala repeat generated a cisacting inhibitory signal interfering with antigen processing and class 1 HLA-restricted presentation. This strategy contributes to the escape from T-cell immunosurveillance by EBV+ tumors with the latency I program of gene expression (i.e., EBNA-1 only), of which BL is the best documented example. Alternatively, the lack of expression of epitopes that are recognized by HLA-restricted, EBV-specific CTL may contribute to the development of EBV+ tumors that feature more extensive gene expression programs (i.e., latency I1 or 111). This situation could be achieved by infection with an EBV genotype that has been adapted to CTL pressure by specific point mutations (De Campos-Lima et al., 1994). Examples of such tumor-specific EBV genotypes are unknown, but the Xhol-loss genotype in Chinese or Alaskan NPC appears to be a good candidate (see earlier). Conversely, the HLA-A2 antigen may confer protection against NPC to Caucasians by virtue of its efficient presentation of LMP-2-derived peptides (Lee et al., 1993; Burt et al., 1994). In conclusion, the studies reviewed have uncovered a wide diversity of EBV genotypes and their relevance for viral epidemiology, transmission, and pathogenesis (Fig. 7). For an uneventful symbiosis of the virus and its host, it is necessary to strike a perfect balance between the various types of EBV-infected cells and the EBV-specific immune response. This need is emphasized by the observation that EBV only causes (malignant) disease in cases that lack EBV-specific immunity, specific cytogenetic accidents, and, possibly, infection with pathogenic viral genotypes.

REFEREKCES

Abdel-Hamid, Mi..Chen, .J. J., Constantine, N., Massoud, hl., and Raab-Traub, N . (1992). I’rrolo&q 190, 168-17.5. Adidingel-. H. K.,Deiius. H.. Freese, U . K., Clark, J . , and Uornkamm, G. U’.(1985). Vzmlogy 141, 22 1-234.

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

247

Aitken, C., Sengupta, S. K., Aedes, C., Moss, D. J., and Sculley, T. B. (1994).]. Gen. Virol. 75,95- 100. Allday, M. J., Crawford, D. H., and Griffin, B. E. (1 988). Nucleic A d Rex 16,4353-4367. Anagnostopoulos, I., Hummel, M., Kreschel, C., and Stein, H. (1995).Blood 85, 744-750. Andersson-Anvret, M., Klein, G., Forsby, N., and Henle, W. (1978). In “Nasopharyngeal Carcinoma: Etiology and Control” (G. de-Tht and Y. Ito, Eds.), pp. 347-357. IARC, Lyon, France. Apolloni, A., and Sculley, T. B. (1994). Virology 202, 978-981. Apolloni, A., Moss, D., S t u p , R., Burrows, S., Suhrbier, A., Misko, I., Schmidt, C., and Sculley, T. (1992). Eur. J. Immunol. 22, 183-189. Arrand, J. R., Young, L. S., and Tugwood, J. D. (1989).J. Virol. 63, 983-986. Baer, R., Bankier, A. T., Biggin, M. D., Deininger, P. L., Farrell, P. J., Gibson, T. J., Hatfull, G., Hudson, G. S., Satchwell. S. C., Sbguin, C., Tuffnell, P. S., and Barrell, B. G. (1984). Nature (London)310, 207-211. Birkenbach, M.,Tong, X.,Bradbury, L. E., Tedder, T. F., and Kieff, E. (1992).J. Exp. Med. 176, 1405-1414. Blacklow, N. R., Watson, B. K., Miller, G., and Jacobson, B. M. (1971).Am.J Med. 51,549552. Bodescot, M., Chambraud, B., Farrell, P.,and Perricaudet, M. (1984). EMBO J. 3, 19131917. Borisch, B., Finke, J., Hennig, I., Delacrttaz, F., Schneider, J., Heitz, P. U., and Laissue, J. A. (1992).J. Pdhol. 168, 229-236. Borisch, B., Hennig, I., Laeng, R. H., Waelti, E. R., Kraft, R., and Laissue,J. (1993a). Blood 82,858-864. Borisch, B., Caioni, M., Hurwitz, N., Dommann-Schemer, C., Odermatt, B., Waelti, E.. Laeng, R. H., Kraft, R., and Laissue, J. (1993b). Inl. J. Cancer 55, 748-752. Bornkamm, G. W., Delius, H., Zimber, U., Hudewentz, J., and Epstein, M. A. (1980).]. Virol. 35, 603-618. Bornkamm, G. W., Hudewentz, J., Freese, U. K., and Zimber, U. (1 982).j. Virol. 43, 952968. Bornkamm, G. W., Von Knebel-Doeberitz, M., and Lenoir, G. M. (1984). Proc. Natl. Acad. SCZ’. USA 81,4930-4934. Bourgeault, I., Gomez, A., Gomard, E., and Levy, J. P. (1991). Clin. Exp. Immunol. 84,501507. Bouzid, M., Belkaid, M. I., Colonna, P., Bouguermouh, A. M., and Ooka, T. (1993). Leukemia 7, 1451-1455. Boyle, M. J., Sewell, W. A,. Sculley, T. B., Apolloni, A., Turner, J. J., Swanson, C. E., Penny, R., and Cooper, D. A. (1991). Blood 78, 3004-3011. Boyle, M. J., Vasak, E., Tschuchnigg, M., Turner, J.J., Sculley, T., Penny, R., Cooper, D. A,, Tindall, B., and Sewell, W. A. (1993). Blood 81,468-474. Brooks, L. A,, Lear, A. L., Young, L. S.. and Rickinson, A. B. (1993).j. Virol. 67, 31823190. Brousset, P., Schlaifer, D., Meggetto, F., Bachmann, E.,Rothenberger, S., Pris, J., Delsol, G., and Knecht, H. (1994). Blood 84,2447-2451. Brown, N.A., Liu, C.-R., Wang, Y.-F., and Garcia, C. R. (1988).J. Virol. 62, 962-969. Buell, P. (1974). CancerRex 34, 1189-1191. Buisson, M., Morand, P., Genoulaz, O., Bourgeat, M. J., Micoud, M.,and Seigneurin, J. M. (1994).J. Cm. Virol. 75, 431-437. Burrows, S. R., Gardner, J., Khanna, R., Steward, T., Moss, D. J., Rodda, S., and Suhrbier, A. (1994).J. Gen. Virol. 75, 2489-2493.

248

IAN W. GRATAMA AND INGEMAR ERNBERG

Burt, K. D., Vaughan, T. L., Nisperos, B., Swanson, M., and Berwick, M. (1994). Int. J. Cancer 56, 465-467. Busson, P., Ganem. G., Flores, P., Mugneret, F., Clausse, B., Caillou, B., Braham, K., Wakasugi, H., Lipinski, M.,and Tursz, T. (1988). Int. J. Cancer 42, 599-606. Cao, S. L., Liu, 71 F., Wang, Z. H., Huang, K. M . , Zhou, J,, Li, D. L., Ni, H . Z., Sang, J. Z., wdng, Y., and Zhang, B. N. (1987). Chin. Oncol. 9, 6-9. Cen, H.. Breinig, M. C., Atchison, R. W.,Ho, M.,and McKnight,J. L. C. (1991).J. Virol. 65, 976-980. Chen, M.L... Tsai. C. N..Liang, C:. L., Shu, C. H., Huang, C. R., Sulitzeanu, D., Liu, S. T., and Chang, Y. S. (1992a). Oncogene 7, 2131-2140. Chen, X., De, V.,Pepper, S., and Arrand, J.R. (1992b).J. Gen. Virol. 73, 463-466. Chen, F., Z o u , J . Z., DiRenzo, L., Winberg, G., Hu, L. F., Klein, E., Klein, E., Klein, G. and Ernberg, 1. (l995).J. ViroE., i7c press. Cho, M . S . , Bornkamin, G. W.,and Zur Hausen, H. (1984).J. Vzrol. 51, 199-207. Choi, P. H. K..Suen, M. W. M.,Huang, D. P., Lo, K. W., and Lee, J. C.K. (1993). Cancer 72, 2873-2878. Cleary, M . L., Nalesnik, ILI. A , Shearer, $4'. T., and Sklar, J. (1988).Blood 72, 349-352. Cohen, J. I . , Wing, F., and Kieff, E. (199l).J. I'iro!. 65, 2545-2554. Countryman, J., and Miller, G. (198.5). Proc. ,Vat/. Acad. Sci. U S A 82, 4085-4089. Countryman, I . , Jenson. H.. Seibl, R., Wolf, H., and Miller, G. (1987).,). Virol. 61, 36723679. Dambaugh. -I.., Hennessy, K.. Chamnankit, L., and Kieff, E. (1984). Proc. Natl. Acad. Sci. U S A 81,763'2-7636. Deacon, E. hi., Pallesen. G.. Niedobitek, G.. Crocker, J., Brooks, L., Rickinson, A. B., and Young. L. S. (1993j.J. Exp. A4ed. 177, 339-349. De Canipos-Linia, P. 0..Gavioli, R., Zhang, Q. J., Wallace, L. E., Dolcetti, R.,Rowe, M., Kickinson, A . B.. and Masucci, M. <;. (1993). Science 260, 98-100. De Campos-Lima. P. O., Lebitsky, V., Brooks, J., Lee, S. P., Hu, L. F., Rickinson, A. B., and Masucci, M.G. (1994).]. Exp. MMed. 179, 1297-1305. De Re. V., Boiocchi, M.,De Vita, S., Dolcetti, R., Gloghini, A,, Uccini, S., Baroni, C;., Scarpa. A., Cattoretti. G., and Carbone, A . (1993). 1nf.J. Cancer 54, 895-898. De Vos, J.F.. and Kuipers, F.H. (1951). S u d . Tijdschr. G r n m k . 95, 3036-3039. Dillner. J., and Kallin, B. (1988).Adu. Cancer Re$. 50, 95-158. Dillner, J., Stern%s.L . , Kallin, B.. Alexander, H., Ehlin-Henriksson. B., Jornvall, H., Klein, G., and Leriier, R. (1984) PI-oc.iVad .4cad. Sri. USA 81, 4652-4656. Dillner, J., Kallin, B.. Klein, G., J6rnvaIl. H., Alexander, H., and Lerner, K.(1985). EMBO 1.4, 1813-1818. Dillner, j..Kallin, B., Ehlin-Henriksson, B.. Rynio. L., Henle, G., Henle, W., and Klein, G. (1986a). I i z t . J . Cancer 37, 19.3-200. Dillnei-.J., Kallin, B.. Alexander, H . . Ernberg, I.. Uno, M., Ono, Y., Klein, G., and Lerner, R. (1986b). Pror. ,Vat/. r l r d . Scz. USA 83, 6641-664.5. Epstein, M.A.. Achong, B. G., and Barr, Y. M. (1964). Lnncet 1, 702-703. Ernberg, I., Kallin, B., Dillner, J., Falk, K., Ehlin-Henriksson, B., Hammarskjdd, M. L., and Klein. G.(1986). In!. 1. Cancer 38, 529-737. Evans, A. S. (1960). '4~1.J. Hyg. 71, 342-362. Falk. K., Ernberg, I., Sakthivel, R., Davis, J., Christensson, B., Luka, J., Okano, M., Grierson, 11. L., Klein, G . , and Purtilo. D. T. (1990). I n t . J. Cancer 46, 976-984. Falk, K., Gratarna. J . b'.,Rowe, M.,Zou, J . Z.. Khanirn, F., Young, L. S., Oosterveer, M .A . P., and Ernberg, I . (199.3).J. Gen. k'ii-ol, 76, 779-790. Farrell, P. J. (19SUj. Ad?'. V i r d Onrol. 8, 103-132.

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

249

Fennewald, S., Van Santen, V., and Kieff, E. (1984).J. Virol. 51, 41 1-419. Finke, J., Rowe, M., Kallin, B., Ernberg, I., RosCn, A,, Dillner, J., and Klein, G. (1987).J. Virol. 61, 3870-3878. Fischer, D. K., Miller, G., Gradoville, L., Heston, L., Weststrate, M. W., Maris, W., Wright, J., Brandsma, J., and Summers, W. C. (1981). Cell 24, 543-553. Fischer, D. K., Robert, M. F., Shedd, D., Summers, W. P., Robinson, J. E., Wolak, J., Stefano, J. E., and Miller, G. (1984). Proc. Natl. Acad. Sci. USA 81, 43-47. Fleischer, G. R., Pasquariello, P. S., Warren, W. S., Zavod, W. S., Korval, A. B., Turner, H. D., and Lennette, E. T. (1981).J. Pediatr. 98, 16-19. Friis, A., Gratarna, J. W., Oosterveer, M. A. P., and Ernberg, I. (1995).J.Med. Virol., inpress. Can, Y. J., Shirley, P., Zeng, Y., and Sixbey, J. W. (1993).J. Infect. Dis. 168, 1349-1355. Gavioli, R., De Campos-Lima, P. O., Kurilla, M. G., Kieff, E., Klein, G., and Masucci, M. G. ( 1992). Proc. Natl. Acad. Sn’. USA 89,5862-5866. Gavioli, R., Kurilla, M. G., De Carnpos-Lima, P. O., Wallace, L. E., Dolcetti, R., Murray, R., Rickinson, A. B., and Masucci, M. G. (1993).J. Virol. 67, 1572-1578. Geertsen, R., Espander-Jansson, A., Dobec, M., Price, P., Wunderli, W., and Ryrno, L. (1994).J. Clin. Microbiol. 32, 112-120. Gerber, P., Walsh, J. H., Rosenblum, E. N., and Purcell, R. H. (1969). Lancet 1, 593-596. Gergely, L., Stern%, L., Dillner, J., and Klein, G. (1984). Intervirology 22, 85-96. Gerhartz, H. H., Mittermuller, j.,Raghavachar, A., Schmetzer, H., Clernrn, C., Kolb, H. J., Bartram, C. C., and Wolf, H. (1988). Int. J. Cancer 42, 672-676. Gilligan, K. J., Rajadurai, P., Lin, J. C., Busson, P., Abdel-Hamid, M., Prasad, U., Tursz, T., and Raab-Traub, N. (1991).J. Virol. 65, 6252-6259. Given, D., and Kieff, E. (1978).J. Virol. 28, 524-542. Gledhill, S., Gallagher, A., Jones, D. B., Krajewski, A. S., Alexander, F. E., Klee, E., Wright, D. H., OBrien, C., Onions, D. E., and Jarrett, R. F. (1991). Br.J. Cancer64,227-232. Goldschmidts, W. L., Bhatia, K., Franklin Johnson, J., Akar, N., GutiCrrez, M. I., Shibata, D., Carolan, M., Levine, A,, and Magrath, I. T. (1992). Leukemia 6, 875-878. Gratama, J. W. (1994).In “Bone Marrow Transplantation” (S.J. Forman, K. G. Blume, and E. D. Thomas, Eds.), 1st ed., pp. 429-442. Blackwell Scientific Publications, Boston, MA. Gratama, J. W., Oosterveer, M. A. P., Zwaan, F. E., Lepoutre, J., Klein, G., and Ernberg, I. (1 988). P roc. Natl. Acad. Sci. USA 85, 8693-8696. Gratama, J. W., Oosterveer, M. A. P., Klein, G., and Ernberg I. (1990a).J. Virol. 64,47034708. Gratama, J. W., Oosterveer, M. A. P., Lepoutre, J. M. M., Van Rood, J. J., Zwaan, F. E., Vossen, J. M. J. J., Kapsenberg, J. G., Richel, D., Klein G., and Ernberg I. (1990b). Transplantation 49, 725-730. Gratama, J. W., Zutter, M. M., Minarovits, J., Oosterveer, M. A. P., Thomas, E. D., Klein, G., and Ernberg, I. (1991). Int.1. Cancer47, 188-192. Gratama, J. W., Lennette, E. T., Lonnqvist, B., Oosterveer, M. A. P., Klein, G., Ringden, O., and Ernberg, I. (1992).J. Med. Virol. 37, 39-42. Gratama, J. W., Oosterveer, M. A. P., Weimar, W., Sintnicolaas, K., Sizoo, W., Bolhuis, R. L. H., and Ernberg, I. (1994).J. Gen. Vzrol. 75, 85-94. Greenspan, J. S., Greenspan, D., Lennette, E. T., Abrams, D. I., Conant, M. A., Petersen, V., and Freese, U. K. (1985). New Eng1.J. Med. 313, 1564-1571. H.arnmerschmidt, W., and Sugden, B. (1989). Nature (London) 340, 393-397. Hanto, D. W., Frizzera, G., Gajl-Peczalska, K. J., Sakamoto, K., Purtilo, D. T., Balfour, H. H., Jr., Simmons, R. L., and Najarian, J. S. (1982).New Eng1.J. Med. 306, 913-918. Hatfull, G., Bankier, A. T., Barrell, B. G., and Farrell, P. J. (1988). Virology 164, 334-340. H(enle, G., and Henle, W. (1970).J. Infect. Dis. 121, 303-310.

250

JAN

W. GRATAMA A N D INGEMAR ERNBERG

Henle, W‘., and Henle, G. (1985).Prog. (;/hi. B i d . Ke.5. 182, 201-209. Henle. G., Henle, M’.,and Diehl, V. (1968). Proc. k n t l . Acad. Sci. L’SA 59, 94-101. Hennessy. K., and Kieff, E. (1983) Proc. LVntl.Acnd. Sri. USA 80, 5665-5669. Hennessy, K., a n d Kieff, E. (1985) Scieircr 227, 1238-1240. Hennessy, K., Heller, M., Van Santen. V.. and Kieff, E. (1983). Science 220, 1396-1398. Hennessy, K., U‘ang, F., Woodland Bushman, E., and Kieff, E. (1986).Proc. N a d Acad. Sci. LISA 83, 9693-5697. Hoagland, R.J. (1955). An<.J . Med. Sci. 229, 262-272. Hu, L. F., Zabarovskv, E. R., Chen, F., Cao, S. L., Ernherg, I . , Klein, G., and Winberg, G. (1991). J . C h i . Viral. 72, 2399-2409. Hu, L. F.. Chen, F., Zheng, X., Ernherg, I . , Cao, S. L., Christensson, B., Klein, G., and Winberg, G. (1993). Oncogene 8, 1575-1583. Hudson. G. S.. Farrell, P. J.. and Barrell, B. G. (1985).J. Virol. 53, 528-535. Hurle); E., and Thorlev-Lawson, D. (1988).J. Ex#. Med. 168, 2059-2075. Imai, S..Koizumi, S., Sugiura, M., Tokunaga, M., Uemura, Y., Yamamoto, N., Tanaka, S., Sato, E., and Osato, T. (19Y4). Pt-or. ~Vatl..4rad. Sct. USA 91, 9131-9135. Israele, V., Shirley, P., and Sixhey, J. W. (1991).J. Irzjecf. Dls. 163, 1341-1343. Jat, P., and Arrand, J. R. (1982).Surleir ’4cidi Res. 10, 3407-3425. Jeng, K. C. G., Hsu, C. Y.. Liu, M. T., Chung, T. T., and Liu, S. T. (1994).J. Clin. Microbial. 32,28-3 I . Jenson, H . B., Farrell, P. J., and Miller, G. (1987).J. Viral. 61, 1495-1506. Joab, I . , Rowe. D. T., Bodescot, M.. Nicolas, J. C., Farrell, P. J., and Perricaudet, M. (1987). J . Viroi. 61, 3340-3344. Joncas, J., a n d kfitnyan. C. (1970). Cat<.Med. Assor. J . 102, 1260-1263. Kallin, B., Dillner, J.. Ernberg, I., Ehlin-Henriksson, B., Rosen, A,, Henle, W., Henle, G . , and Klein, G. (1986). Proc. ‘Vutl. .4cad. Sri. USA 83, 1499-1503. Kamel. 0. W..Van d e Rijn, M., Hanasono, Xi. M., and Warnke, R. A. (1995).Leukemia Ljmphoma 16, 363-368. Karran, L., Gao, Y., Smith, P. R., and Griffin, B. E. (1992). Proc. Natl. Acud. Sci. USA 89, 8058-8062. Katz, B. Z., Andiman, M! .4., Eastman, R., Martin, K., and Miller, G. (1986).J. Infect. Dis. 153,601-604. Katz, R . 2.. Niederman, J. C., Olson, B. A , and Miller, G. ( 1 9 8 8 ) .j.Infect. Dis. 157, 299508. Kaye, K . M., Izumi, K. M., Mosialos, G., and Kieff, E. (1995).J. Virol. 69, 675-683. Kempkes, B., Spitkovsky, D., Jansen-Dlrr, P., Ellwart, J. W., Kremmer, E., Delecluse, H. J., Rottenberger, C . , Bornkamni, G. W., and Hammerschmidt, W. (1995). EMBO J. 14, 88-96. Kerdiles, B . , Walls. D., Triki, H.. Perricaudet, M., and Joab, I. (l990).J. Virol. 64, 18121816. Kerr, B. M..Leal-, A. L.. Rowe, M . . Croom-Carter, D., Young, L. S.. Rookes, S. M., Gallimore, P. H., and Rickinson, A. B. (1992). Viroloa 187, 189-201. Khanna, R., Burrows. S. R.. Kurilla, M. G.. Jacob, C. A., Misko, I. S., Sculley, T. B., Kieff, E., and Moss. D. J . (1992).J . Exp. M p d . 176, 169- 176. Kieff, E., and Liehowitz, D. (1990). In “Virology” (B. Fields, D. Knipe, R. Chanock, M. Hirsch. a n d J. Melnick, Eds.), 2nd ed.. pp. 1889-1920. Raven Press, New York. Klein, G . (1989).Cell 58, 5-8. Klein, G., and Dombos, L. (1973). Int. J . Cnticer 11, 327-337. Knecht. H.. Bachmann, E., Joske, D. J. L., Sahli, R., Emery-Goodman, A , , Casanova, J. L., Zilii.. 51.. Bachniann, F.. and Odermatt. B. F. (1993a). Leukemin 7 , 580-585.

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

25 1

Knecht, H., Bachmann, E., Brousset, P., Sandvej, K., Nadal, D., Bachmann, F., Odermatt, B. F., Delsol, G., and Pallesen, G. (1993b).Blood 82, 2937-2942. Knecht, H., Martius, F., Bachmann, E., Hoffman, T., Zimmermann, D. R., Rothenberger, S., Sandvej, K., Wegmann, W., Hurwitz, N., Odermatt, B. F., Kummer, H., and Pallesen, G. (1995). Leukemia 9, 458-465. Krishnamurthy, S., Lanier, A. P., Dohan, P., and Lanier, J. F. (1987).Hum. Pathol. 18,986996. Kunimoto, M., Tamura, S., Tabata, T., and Yoshie, 0. (1992).J. Gen. Virol. 73, 455-461. Kyaw, M. T., Hurren, L., Evans, L., Moss, D. J., Cooper, D. A., Benson, E., Esmore, D., and Sculley, T. B. (1992). AIDS Res. Hum. Retrouiruses 8, 1869-1874. Kyaw-Tanner, M. T., Esmore, D., Burrows, S. R., Benson, E. M., and Sculley, T. B. (1994). Transplantation 57, 1611-1617. L,abrecque, L. G., Barnes, D. M., Fentiman, I. S., and Griffin, B. E. (1995). Cancer Res. 55, 39-45. L,aux, G., Perricaudet, M., and Farrell, P. J. (1988).E M B O J . 7, 769-774. L,ee, S. P., Thomas, W. A,, Murray, R. J., Khanim, F., Kaur, S., Young, L. S., Rowe, M., Kurilla, M., and Rickinson, A. B. (1993).J. Virol. 67, 7428-7435. L.ee, E. S., Locker, J., Nalesnik, M., Reyes, J., Jaffe, R., Alashari, M., Nour, B., Tzakis, A., and Dickman, P. S. (1995). New Engl. J. Med. 332, 19-25. L,ee, S. P., Morgan, S., Skinner, J., Thomas, W. A,, Rowland Jones, S., Sutton, J., Khanna, R., Whittle, H. C., and Rickinson, A. B. (1995). Eur.J. Immunol. 25, 102-110. L.ees,J. F., Arrand, J. E., De, V., Pepper, S., Stewart, J. P., Mackett, M., and Arrand, J. R. (1993). Virology 195, 578-586. L,erner, M. R., Andrews, N. C., Miller, G., and Steitz, J. A. (1981).Proc. Natl. Acad. Sci. USA 78, 805-809. L.evitskaya,J., Coram, M., Levitsky, V., Steigerwald-Mullen, P. M., Klein, G., Kurilla, M. G., and Masucci, M. G. (1995). Nature, in press. L,eyvraz, S., Henle, W., Chahinian, A. P., Perlman, C., Klein, G., Gordon, R. E., Rosenblum, M., and Holland, J. F. (1985). New Engl. J. Med. 312, 1296-1299. L.i, S. L., Biberfeld, P., and Ernberg, I. (1994).Int.1. Cancer 59, 287-295. Liebowitz, D. (1995).New Engl. J. Med. 332, 55-57. Liebowitz, D., Wang, D., and Kieff, E. (1986).J. Virol. 58, 233-237. Lain,J. C., De, B. K., and Lin, S. C. (1993a1.1. Viral. Methods 43, 233-246. Lin, J. C., Lin, S. C., De, B. K., Chan, W. C., and Evatt, B. L. (1993b). Blood 81, 3372338 1 . L,in, J. C., Lin, S. C., Luppi, M., Torelli, G., and Mar, E. C. (1995).J.Med. Virol. 45, 183191. L,ongnecker, R., and Kieff, E. (1990).J . Virol. 64, 2319-2326. L.ung, M. L., and Chang, G. C. (1992). Int. J. Cancer 52, 34-37. L.ung, M. L., Chang, R. S., and Jones, J. H. (1988).J. Virol. 62, 3862-3866. L,ung, M. L., Chang, R. S., Huang, M. L., Guo, H. Y., Choy, D., Sham, J., Tsao, S. Y., Cheng, P., and Ng, M. H. (1990). Virology 177, 44-53. L.ung, M. L., Lam, W. P., Sham, J., Choy, D., Zong, Y. S., Guo, H. Y., and Ng, M. H. (1991). Virology 185, 67-7 1 . L.ung, M. L., Lam, W. P., Chan, K. H., Li, S., Sham, J., and Choy, D. (1992).Int.J. Cancer 52, 174- 177. Lung, M. L., Chang, G. C., Miller, T. R., Wara, W. M., and Phillips, T. L. (1994). Int. J. Cancer 59, 743-746. Magrath, I. (1990). Adv. Cancer Res. 55, 133-270. Masucci, M. G., and Ernberg, I. (1994). Trends Microbiol. 2, 125-130.

252

J A N W. GRATAMA A N D INGEMAR ERNBERG

Masucci, M .G . , Gavioli, R., De Campos-Lima, P. O., Zhang, Q.J., Trivedi, P., and Dolcetti, R. (1993). A n n . ”1’. Acad. Sci. 690P, 86-100. McClain, K. K., Leach, C. T...Jenson. H. B., Joshi, V. V., Pollock, B. H., Parmley, R. T., DiCarlo, F. J . , Gould Chadwick. E., and Murphy, S. B. (1995). New Engl. J . Med. 332, 12-18. Miller, G. (1990).J. Infect. Dis. 161, 833-844. Miller, G., Shope, T., Lisco, H., Stitt, D., and Lipman, M. (1972). Proc. Natl. Acnd. Sci. USA 69, 383-387. Miller, G., Robinson, J., Heston, L., and Lipman, M. (1974). Proc. Natl. Acad. Scz. USA 71, 4006-40 10. Miller. G . , Rabson, M., and Heston, L. (1984).J. Virol. 50, 174-182. Miller, G., Heston, L., and Countryman, J. (1985a).J. Vzrol. 54, 45-52, Miller, G., Grogan, E.. Fischer, D. K., Niederman, J. C., Schooley, R. T., Henle, W., Lenoir, G.. and Liu, C. R. (l985b).New Engl. J. Med. 312, 750-755. Miller, W.E., Hood Edwards, R., Walling, D. M., and Raab-Traub, N.(1994).J.Gen. Virol. 75,2729-2740. Misko, I . S, Pope, J. H., Hutter, R., Soszynski, T. D., and Kane, R. G. (1984).Int. J . Cancer 33,239-243. Moorthy, R. K., and Thorley-Lawson, D.A. (1993). J . Virol. 67, 2637-2645. Morgan, D. G . , Niederman, J. C., Miller, G., Smith, H. W., and Dowaliby, J. M. (1979). Lancet 2, 11.54-1 157. Moss, D., Rickinson, A,. and Pope, J. (1978). Int. J . Cancer 22, 662-668. Moss, D. J . , Misko, I. S., Burrows, S. K.,Burman, K., McCarthy, R., and Sculley, T. B. (1988). N a t u r ~(London) 331, 719-721. Muller-Lantzsch. N.,Lenoir, G., Sauter. M., Takaki, L., Bechet, J. M., Kuklik-Roos, C., Wunderlich. D., and Bornkamm, G . (1985). E M B O J . 4, 1805-1811. Murray, R. J . , Kurilla. M. G., Brooks, J. hi., Thomas, W. A., Rowe. M., Kieff, E., and Rickinson. A . B. (1992).J.Exp. Med. 176, 157-168. Nemerow, C. R., Moore, M. D., and Cooper, N. R. (1990). .4du. Cancer Res. 54, 273-300. Niederman, J. C., Evans, A. S., Subrahmanyan. L., and McCollum, R. W. (1970). NewEngl. J . M d 282, 361-365. Nilsson, K . , Klein, G., Henle, W.. and Henle, G. (1971). Int. J. Cancer 8, 443-450. Oosterveer, M. A. P., Markusse, H. M., Lennette, E. T., Zou, J. Z., Bolhuis, R. L. H., and Graiama, J. W. (1993).J . Med. Virol. 41, 261-269. Pallesen, G., Hamilton-Dutoit, S. J., Rowe, M., and Young, L. S. (1991). Lancet 337, 320322. Pallesen, G . , Hamilton-Dutoit, S. J., and Zhou. X. (1993). Ad7~Cancer Res. 62, 179-239. Papadopoulos, E. B., Ladanyi, M., Emanuel, D., Mackinnon, S., Boulad, F., Carabasi, M. H., Castro-hialaspina, H., Childs, B. H., Gillio, A. P., Small, T. N., Young, J. W., Kernan, N . , and O’Reilly, R. J. (1994). Neul Engl.J. Med. 330, 1185-1191. Patton, D. F., Shirley, P., Raab-Traub, N., Resnick, L., and Sixbey, J- W. (I990).J. Virol. 64, 397-400. Patton, D. F., Kibeiro, R. C., Jenkins, J. J., and Sixbey, J . W. (1994).J.Infect. Dts. 170, 7-12. Perricaudet, M.. Akusjarvi, G., Virtanen, A , , and Petterson, U. (1979).Nature (London) 281, 694-696. Petti, L., and Kieff, E. (I988).J. Vtrol. 62, 2173-2178. Petti, L., Sample. J . , Wang, F., and Kieff, E. (1988). J. Vzrol. 62, 1330-1338. Pizzo, P. A , , Magrath, I. T., Chattopadhyady, S. K., Biggar, R. J . , and Gerber, P. (1978). ,Vature (London) 272, 629-63 1.

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

253

Polack, A., Hartl, G., Zimber, U., Freese, U. K., Laux, G., Takaki, K., Hohn, B., Gissman, L., and Bornkamm, G. W. (1984). Gene 27, 279-288. Pulvertaft, R. J. V. (1964). Lancet 1, 238-240. Pulvertaft, R. J. V. (1965).J. Clin. Pathol. 18, 261-273. Qu, L., and Rowe, D. T. (1992).J. Virol. 66, 3715-3724. Raab-Traub, N., and Flynn, K. (1986). Cell 47, 883-889. Raab-Traub, N., Dambaugh, T., and Kieff, E. (1980). Cell 22, 257-267. Rabson, M., Gradoville, L., Heston, L., and Miller, G. (1982).J. Vzrol. 44, 834-844. Rabson, M., Heston, L., and Miller, G. (1983). Proc. Natl. Acad. Sci. USA 80, 2762-2766. Reedman, B. M., and Klein, G. (1973). Int.J. Cancer 11, 499-520. Rickinson, A. B., Finerty, S., and Epstein, M. A. (1977). Int. J. Cancer 19, 775-782. Rickinson, A. B., Moss, D. J., Wallace, L. E., Rowe, M., Misko, I. S., Epstein, M. A,, and Pope, J. H. (1981). Cancer Res. 41, 4216-4221. Rickinson, A. B., Yao, Q. Y., and Wallace, L. E. (1985). Br. Med. Bull. 41, 75-79. Rickinson, A. B., Young, L. S., and Rowe, M. (1987).J. Virol. 61, 1310-1317. Ricksten, A., Kallin, B., Alexander, H., Dillner, J., Fihraeus, R., Klein, G., Lerner, R., and Rymo, L. (1988). Proc. Natl. Acad. Sci. USA 85, 995-999. R’occhi, G., Hewetson, J., and Henle, W. (1973). Int. J. Cancer 11, 637-647. R’ogers, R. P., Strominger, J. L., and Speck, S. H. (1992). Adv. Cancer Res. 58, 1-26. R’ooney,C., Taylor, N., Countryman, J., Jenson, H., Kolman, J., and Miller, G. (1988). Proc. Natl. Acad. Sci. USA 85, 9801-9805. Rowe, D. T., and Clarke, J. (1989).J. Gen. Vzrol. 70, 1217-1229. Rmowe, M., Rowe, D. T., Gregory, C. D., Young, L. S., Farrell, P. J., Rupani, H., and Rickinson, A. G. (1987a). EMBOJ. 6, 2743-2751. Rsowe, M., Evans, H. S., Young, L. S., Hennessy, K., Kieff, E., and Rickinson, A. B. (198713). J. Gen. Vzrol. 68, 1575-1586. Rsowe M., Young, L. S., Cadwallader, K., Petti, L., Kieff, E., and Rickinson, A. B. (1989).J. Virol. 63, 1031-1039. Rowe, M., Lear, A. L., Croom-Carter, D., Davies, A. H., and Rickinson, A. B. (1992).J. Vzrol. 66, 122-131. Rymo, L., Klein, G., and Ricksten, A. (1985). Proc. Natl. Acad. Sci. USA 82, 3435-3439. Sample, J., Humrnel, M., Braun, D., Birkenbach, M., and Kieff, E. (1986). Proc. Nutl. Acud. Sci. USA 83,5096-5100. Sample, J., Liebowitz, D., and Kieff, E. (1989).J. Virol. 63, 933-937. Sample, J., Young, L., Martin, B., Chatman, T., Kieff, E., Rickinson, A., and Kieff, E. (1990).J. Virol. 64, 4084-4092. Sample, J., Brooks, L., Sample, C., Young, L., Rowe, M., Gregory, C., Rickinson, A. B., and Kieff, E. (1991). Proc. Natl. Acad. Sci. USA 88, 6343-6347. Sample, J., Kieff, E. F., and Kieff, E. D. (1994).J. Gen. Virol. 75, 2741-2746. Sandvej, K., Peh, S. C., Storstein Andresen, B., and Pallesen, G. (1994). Blood 84, 40534060. Sawyer, R. N., Evans, A. S., Niederman, J. C., and McCollum, R. W. (1971).J. Infect. Dis. 123, 263-270. Schaefer, B., and Speck, S. (1994). In “Epstein-Barr Virus & Associated Diseases” (J. Pagano and A. Rickinson, Eds.), p. 192. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Schaefer, B., C., Woisetschlaeger, M., Strominger, J. L., and Speck, S. H. (1991). Proc. Nutl. Acad. Scz. USA 88, 6550-6554. Sculley, T. B., Spelsberg, T. C., and Pearson, G. R. (1984). Intervirology 22, 191-200.

"4

JAN W. GRATAMA AND INGEMAR ERNBERG

Sculley. T. B., Moss, D. ,J., Hazelton, R. A , . and Pope. J. H. (I987)J. Ckn. Vzrol. 68,20692078. Sculley, 7'. B., Sculley. D. C;., Pope, J. H., Bornkanini, G. W., Lenoir, G., and Rickinson, A . B. (1988a). IntenJirolo,g 29, 77-85. Sculley, T. €3.. Cross, S. XI., Borrow. P., and Cooper. D. A . (1988b).J. Infect. DZS. 158, 186192.

Sculley, T. B., Apolloni, A , , Stumm, R., hloss. D. J., Miiller-Lantzsch, N., Misko, I. S., and Coopei-. D. A. (1989).I'irolua 171, 401-408. Sculley, T. B.. Apolloni. A , , Hurren, L.. Moss, D.J., and Cooper, D. A. (1990).J. Infect. Dis. 162,643-648, Shibata. D..and Weiss, L. M. (1992).Am. J . Pathof. 140, 769-774. Shibata, D., Tokunaga, M.,Uemura, Y., Sato, E., Tanaka, S.,and Weiss, L. M. (1991).Am.J . Patliol. 139, 169-474. Shiniizu, N., E'amaki. hl., Sakuma, S., Ono, Y.. and Takada. K. (1988). Int. J . Cancer 41, 7.14-75 1. Shu, C. H.. Chang, Y. S.. Liang, <:, L.. Liu. S. T., Lin, C. 2.. and Chang, P. (l992).J. V i d .\fethod.y 38, 123- 130. Sixbey, 1. \V., and Yao, Q. Y. (1992). Srimre 255, 1578-1580. Sixbev, J. \V., Nedrud. J. G., Raab-Traub, S . , Hanes, R. A., and Pagano, J. S. (1984). New Eltgl. J . ,M&. 310, 122.5-1230. Sixbev, J . W., Lemon, S. M.,and Pagano, J. S. (1986). t m u 2~, 1122-1 124. Sixbey, J. M:, Shirley. P., Chesne); P. J.. Buntin, D. M.,and Resnick, L. (1989). Laiiret 2, 761-565. Sixbey. J. W.. Shirley. P.. Sloas, Sf.. Raab-Traub. N., and Israele, V. (1991).J. Infect. Dis. 163, 1008- I 0 15. Strnad, B.C., Schuster, T. C . , I-Iopkins. R. F., 111, Seubaiier, R. H., and Rabin, H . (1981).J. 1,'iroI. 38, 996- 1004. Sugden, B.. Phelps. hi.. and Donioradzki, J. (1979).J. V m l . 31, 390-395. S u n i a y , C. V., and Ench, Y. (1986)./. /?tfect. OU. 154, 842-850. Summers. M'. P., Grogan, E. A , , Shedd, D., Robert, M.,Liu, C:. R..and Miller, G. (1982). P w c . . V d . Arad. Sri. USA 79, 5688-5692. Tao. Q.. Srivastava, G.. Chan, A. C . L., Chung, L. P., Loke, S. L.. and Ho, F. C. S. (1995).J. M e d . l'irol, 45, 71-77, Tdyk)r, K . .4.. Wetzel. S., Lyles, D. S.. and Pollok. P. A. (I994a).J. l ' i d 68, 6421-6431. Taylor, Y.. Melvin, W.T., Sewell. H. F., Flannelly, G., ; t ~ t t lWalker, F. (1994b).J. C h . Pathol. 47, 92-YX Thomas. J. A . , .-\llday. Sf.,J., and Crawft,rtl. I). H. (1991). Ad?'. C a r ~ e rRe$. 57, 329-380. 'T'ieiney, R. J.. Steven, N..Young, L. S.. and Rickinson, A. B. (1994).J . V i m / . 68,7374'7385. Trivedi, P., Hu, L. F., Chen, F., Christensson, B., Masucri, XI.. Klein, G.. and Winberg, G. (19%). Eio-. J . Chttcer 30A, 84-88. Turner. A . R.. %I(Donald.R. N., and Cooper. B. A. (1972). " t i / ) t . I n t m . M e d . 77, 751-753. \'an Gelder. T.. Kroes. L. C:. %l., 3lulcier, A . , Grarania, J - W.,and Weiniar, W. (1994). 7rnri.\pln,itntio,r 58, 852 -85 5. Van Santen. I,.,(:heung. A , , and Kief'f, E. (1981).Pror. .lid..4cud. Srz. USA 78, 1930-1934. M'ahren. B.. l a i t o r p , K . . Sterner, G,and Espniark, .A. (1970). Proc.. Sor. Exp. B i d . Med. 133, 934-939. Wallace. L. t..Kirkinson, A . B., Rotre. M,, and Epstein. M.A.(1982). .l'ature 297, 413-415. Walling. D. hl., a t i d Raab-Traub, N. (1994)./. I'irof. 68, 7909-7917.

MOLECULAR EPIDEMIOLOGY OF EBV INFECTION

255

Walling, D. M., Neese Edmiston, S., Sixbey,J. W., Abdel-Hamid, M., Resnick, L., and RaabTraub, N. (1992). Proc. Natl. Acad. Sci. USA 89, 6560-6564. Walling, D. M., Perkins, A. G., Webster-Cyriaque, J., Resnick, L., and Raab-Traub, N. ( 1994).J. Virol. 6 8 , 79 18-7926. Wang, F., Petti, L., Braun, D., Seung, S., and Kieff, E. (1987).J. Viral. 6 1 , 945-954. Weiss, L. M., Movahed, L. A., Butler, A. E., Swanson, S. A., Frierson, H. F., Jr., Cooper, P. H., Colby, T. V., and Mills, S. E. (1989). Am.]. Surg. Pathol. 11, 625-631. Wising, P. J. (1942). Acta Med. S c a d . 109, 507-512. Woisetschlager, M., Strominger, J. L., and Speck, S. H. (1989).Proc. Natl. Acad. Sci. USA 8 6 , 6498-6502. Woisetschlager, M., Yandava, C. N., Furmanski, L. A., Strominger, J. L., and Speck, S. H. (1990). Proc. Natl. Acad. Sci. USA 87, 1725-1729. Yaio, Q. Y., Ogan, P., Rowe, M., Wood, M., and Rickinson, A. B. (1989).Int. J. Cancer43, 6771. Yao, Q. Y., Rowe, M., Martin, B., Young, L. S., and Rickinson, A. B. (1991).J. Gen. Viral. 7 2 , 1579- 1590. Yates, J. L., and Guan, N. (1991).J. Virol. 6 5 , 483-488. Young, L. S., Yao, Q. Y., Rooney, C. M., Sculley, T. B., Moss, D. J., Rupani, H., Laux, G., Bornkamm, G. W., and Rickinson, A. B. (1987).J. Gen. Viral. 68,2853-2862. Young, L. S., Dawson, C. W., Clark, D., Rupani, H., Busson, P., Tursz, T.,Johnson, A,, and Rickinson, A. B. (1988).J. Gen. Viral. 69, 1051-1065. Young, L. S., Alfieri, C., Hennessy, K., Evans, H., OHara, C., Anderson, K. C., Ritz, J., Shapiro, R. S., Rickinson, A., Kieff, E., and Cohen, J. I. (1989). New Engl. J. Med. 321, 1080- 1085. Zimber, U., Adldinger, H. K., Lenoir, G. M., Vuillaume, M., Knebel-Doeberitz, M. V., Laux, G., Dtsgranges, C., Wittmann, P., Freese, U. K., Schneider, U., and Bornkamm, G. W. (1986). Virology 154, 56-66. Ziir Hausen, H., Schulte-Holthausen, H. S., Klein, G., Henle, W., Henle, G., Clifford, P., and Santesson, L. (1970). Nature (London) 228, 1056-1058.

This Page Intentionally Left Blank

SCATTER FACTOR AND ANGIOGENESIS Eliot M. Rosen and ltzhak D. Goldberg Department of Radiation Oncology, Long Island Jewish Medical Center, The Long Island Campus for Albert Einstein College of Medicine, New Hyde Park, New York 11042

I. Introduction: Scatter Factor (Hepatocyte Growth Factor) and the c-met Receptor A. Scatter Factor (SF) B. c-met Receptor C. Biologic Activities of SF 11. SF Biologic Actions on Blood Vessel Wall Cells in Vitro and in Vivo A. Vascular Endothelial Cells (ECs) B. Vascular Smooth Muscle Cells (SMCs) and Pericytes C. In Vivo Angiogenic Activity 111. SF as a Potential Tumor Angiogenesis Factor A. Angiogenesis in Human Cancers B. Expression of SF within Tumors C. Mechanisms of Tumoral SF Production IV. Role of SF in Angiogenesis: Hypotheses and Future Directions A. Linkage of Angiogenesis and Tumor Suppressors B. SF- and c-met-Inducing Factors C. Directions for Further Experimental Study D. Potential Clinical Applications V. Summary and Conclusions References

I. Introduction: Scatter Factor (Hepatocyte Growth Factor) and the c-met Receptor

A. SCATTERFACTOR(SF) SF is a mesenchyme-derived protein that dissociates (“scatters”)sheets of epithelium (Stoker et al., 1987; Rosen et al., 1989). SF is identical to hepatocyte growth factor (HGF) (Weidner et al., 1991a; Bhargava et al., 1992), a serum mitogen for rat hepatocytes that is thought to function as a m hepatotrophic factor for liver repair (Miyazawa et al., 1989; Nakamura et al., 1989). SF is a heparin-binding glycoprotein consisting of a 60-kDa a-chain and a 30-kDa P-chain (Gherardi et al., 1989; Rosen et al., 1990b; Weidner et al., 1990). The a-chain contains an N-terminal hairpin loop and four “kringles” (disulfide looped structures that mediate protein-protein interactions). The P-chain is homologous to serine proteases. SF has 38% amino acid sequence identity to the proenzyme plas257 ADVANCES 1N CANCER RESEARCH, VOL. 67

Copyright 0 1995 by Academic Press, Inc Ail rights of I-eproduction in any form reserved

258

ELIOT M. ROSEN AND ITZHAK D. COLDBEKG

minogen (h'akamura et al., 1989), but lacks protease activity (Rosen et al., 1990b) due to the replacement of two essential amino acids at the catalytic triad of the P-chain. SF is synthesized as a 728-amino-acid precursor (preproSF); intracellular cleavage of a 3 1-amino-acid signal peptide results in its secreted single-chain form (proSF), which is biologically inactive (Lokker et al., 1992). Extracellular cleavage of proSF at 494Arg4""Val yields active two-chain SF. HGF activator, a novel serine protease homologous to coagulation factor XI1 (Hagemann factor), may be a physiologic cleavage enzyme for SF (Miyazawa et al., 1993). This enzyme is produced in zymogen form; it may be activated by a proteolytic cascade initiated by tissue injury (Miyazawa et al., 1994). Plasminogen activators (uPA and tPA) can also cleave and activate proSF, but only at supraphysiological concentrations (Naldini et al., 1992; Mars et al., 1993). B. c-met RECEPTOR

The SF receptor is the protein product of the c-met protooncogene (Bottaro et al., 1991), a transmembrane tyrosine kinase (TK) expressed predominantly by epithclia (Gonzatti-Haces el al., 1988). T h e c-met receptor is a 190-kDa glycoprotein consisting of a 145-kDa membranespanning P-chain and a 50-kDa a-chain that is expressed on the cell s u r f x T h e extracellular binding, transmembrane, and intracellular kinase, 2nd noncatalytic phosphate acceptor domains are located on the @chain. Studies suggest that much of the signal transduction from the SF-activated c-met receptor occurs through the interaction of a novel tanc!em YV(H/IV)V motif with the src homology-:! (SH2) domains of various intracellular signaling molecules (Ponzetto et al., 1994). Tyrosine phosphorylation at this site mediates the binding of c-met to phosphatidylinositoM'-kinase, protein tyrosine phosphatase 2, phospholipase Cy, ppCio'-
C. BIOLOGICACTIVITIES OF SF SF transduces three major classes of cellular actions in vitro: motility, growth, and morphogenesis. Studies employing chimeric receptor constructs indicate that each of these actions can be transduced by the c-met

SCATTER FACTOR AND ANGIOGENESIS

259

T K (Weidner et al., 1993; Zhu et al., 1994). In addition to cell dissociation, SF induces random movement of isolated epithelial cells, chemotactic (gradient-directed) migration, migration from carrier beads to flat surfaces, and invasion through extracellular matrix proteins (Rosen et al., 1990b,c, 1991a; Weidner et al., 1990; Bhargava et al., 1992; Li et al., 1!394). SF stimulates the mRNA and protein expression of both uPA and uPA receptor (uPAR) (Pepper et al., 1992; Grant et al., 1993; Rosen et al., 1!394b). T h e net effect is to increase the amount of uPA bound to uPAR 011 the cell surface. Receptor-bound uPA on the cell surface is catalytically active and is thought to mediate focal degradation of the extracellular matrix necessary to clear a path for invading cells (Saksela and Rifkin, 1988). Thus, SF appears to be able to “switch on” a program of cell activities for invasion. T h e switching mechanism may involve inductiton of an as yet unidentified transcription factor. SF is mitogenic for various normal cell types, including epithelial cells, vascular endothelial cells, and melanocytes (Kan et al., 1991; Rubin et al., 1991; Halaban et al., 1‘392).SF is also a potent morphogen. SF induces MDCK epithelial cells incubated in collagen I gels to organize into a network of branching tubules with the proper apical-basolateral polarity (Montesano et al., 1’991;Santos et al., 1993). Similarly, SF induces mammary epithelial cells to form ductlike structures (Tsarfaty et al., 1992). Thus, SF can activate specific programs of cell differentiation depending upon the cell type and environment. 11. SF Biologic Actions on Blood Vessel Wall Cells in Vitro and in Vivo A,. VASCULAR ENDOTHELIAL CELLS(ECs)

1. EC as a SF Producer Cell Vascular ECs may function both as a producer of SF and as a target of SF action. ECs in vitro and in vivo normally produce little or no SF (Stoker et al., 1987; Rosen et al., 1989; Matsumoto et al., 1993). However, following liver injury, hepatic sinusoidal ECs as well as pulmonary alveolar ECs synthesize SF, as demonstrated by in situ hybridization to detect SF-specific mRNA transcripts (Yanagita et al., 1992; Noji et al., 1990). These findings suggest that appropriately stimulated ECs are capable of producing SF. They further suggest that paracrine and endocrine signals associated with tissue injury are responsible for the induction of EC synthesis of SF. Some of the putative factors that may transmit such a signal are discussed later (Section III.C.3).

260

ELIOT M . ROSEN A N D ITZHAK D. COLDBERG

2.EC as an SF Target Cell During the early stages of angiogenesis i.n uivo, ECs from preexisting small vessels (usually venules that lack a smooth muscle covering) focally degrade the subendothelial basement membrane, migrate out into the interstitium toward an angiogenic stimulus, and form capillary sprouts (Folkman, 1983). Sprouting ECs proximal to the migrating tip proliferate; subsequently, the EC sprouts organize into an anastamosing network of capillary tubes. Finally, these ECs synthesize new basement membrane, Adhesion of SMCs and pericytes and formation of new basement membrane are processes associated with the termination of the angiogenic response (Folkman, 1985; Antonelli-Orlidge et al., 1989). Stimulation of EC motility, proliferation, and capillary-like tube formation in vitro are thought to correlate with the ability to induce angiogenesis in zu’zio, since each of these processes occurs during the formation of new blood vessels (Folkman, 1985). Both large vessel- and microvessel-derived ECs express the c-met receptor and are biologically responsive to SF (Rosen et al., 1990b,c, 1991b; Bussolino et al., 1992; Grant et al., 1993; Naidu et al., 1994). SF is chemotactic to ECs and stimulates random motility, as demonstrated in assays using microwell modified Boyden chambers (Rosen et al., 1990b, 1991b). In addition, SF induces the migration of ECs cultured on microcarrier beads from the beads to flat culture surfaces (Rosen et al., 1990b,c). In chenioinvasion assays, a gradient of SF induces the penetration of ECs through porous filters coated with Matrigel, a reconstituted matrix of basement membrane (Rosen et al., 1991b). Maximal chemotaxis, bead migration, and invasion of human umbilical vein ECs (HUVEC), calf pulmonary artery ECs (CPAE), bovine aortic ECs (BAEC), and bovine brain ECs (BBEC) are typically observed at SF concentrations of 2-20 ng/ml. In studies using the microcarrier bead migration assay, we found that BBEC migration was stimulated 5-fold by SF, but was unaffected by basic FGF or EGF (Rosen et al., 1991~).On the other hand, TGFS blocked both basal and SF-stimulated migration of BBEC. Migration from carrier beads was blocked by inhibitors of protein synthesis (cycloheximide), but not by inhibitors of DKA synthesis (hydroxyurea) (Rosen rf al., l99lc). In addition to motility, SF stimulates DNA synthesis and proliferation of some EC types, including HUVEC and human omental microvessel ECs (Rubin et al., 1991; Morimoto et al., 1991). Capillary tube formation appears to be an independent property of ECs not directly related to motility or proliferation (Grant et al., 1989). When ECs are plated onto a surface of reconstituted basement membrane (Matrigel), they cease

SCATTER FACTOR AND ANGIOGENESIS

26 1

DNA synthesis and proliferation, extend long cytoplasmic processes, and begin to organize into a network of capillary-like tubes. SF stimulates capillary-like tube formation in HUVEC and BBEC cultures by up to 5-lO-fold, as determined by computerized digital image analysis of stained cultures (Rosen et al., 1991b; Grant et al., 1993). SF also induces large increases in the' expression of uPA activity by EC cultures (Rosen et al., 1991b; Grant et al., 1993). Most of the SF-induced uPA activity is cell-associated rather than secreted. The majority of cellassociated uPA is bound to uPA receptor on the cell surface, where it is well-positioned to mediate focal degradation of extracellular matrix proteins, a prerequisite for invasion (Saksela and Rifkin, 1988). Taken together, these findings indicate that SF can induce most or all of the phenotypic characteristics expected of ECs undergoing angiogenesis (illustrated in Fig. l). B. VASCULAR SMOOTH MUSCLECELLS(SMCs) AND PERICYTES

1. SMC as a SF Producer Cell In vitro, bovine aortic, human iliac artery, and rat arterial SMCs produce SF at rates comparable to those of high producer human lung fibroblasts (e.g., MRC5, WI38) (64-128 scatter unitdl06 cells/48 hr) (Rosen et al., 1989, 1990a). The biological and chemical properties of SMC-derived SF are very similar to those of fibroblast-derived SF, and it is likely that these molecules are identical (Rosen et al., 1990a,b). 2. SMC as an SF Target Cell

Psoriasis is a chronic inflammatory skin disease characterized by the proliferation of epidermal keratinocytes and neovascularization in dermal papillae and papillary dermis. Cells of the microvessel wall (pericytes, ECs) in psoriatic plaques stain positively for c-met protein (Grant et al., 1993), suggesting that these cell types are potential target cells for SF. Smooth muscle cells in tumor microvasculature also express high levels of immunoreactive SF (Section 1V.B). Recruitment of SMCs and pericytes (which are generally regarded as microvascular SMCs) is an essential component of angiogenesis. These cells are thought to stabilize newly formed vessels, thus contributing to termination of the angiogenic process (Antonelli-Orlidge et al., 1989). Therefore, it seems logical that SF itself might induce the influx and/or proliferation of these cells at the appropriate time during angiogenic response. Cultured pericytes from bovine retina express c-met mRNA, as de-

262

..

ELIOT M . ROSEN AND ITZHAK D. GOLDBERG

A

0.0

-1

w

04 0 0 ~

9 200

??

I

0

0.5

2.5

I

B

1

10.0

-

T

0.0 0.2 2.0 10.0 20.0

0.2

2.0 20.0

D 60

E E f 40 v

P

9 W

m 20 3 I-

0 0.0

2.0 20.0 200.0

SCAlTER FACTOR CONCENTRATION (nglml)

FIG. 1. Angiogenic endothelial cell (EC) phenotype induced by SF. EC chemotaxis, protease production, invasion, and capillary-like tube formation, processes required during angiogenesis, are each stimulated by SF. (A) Chemotaxis. CPAEs were induced to migrate across collagen-coated 8-pm Nucleopore filters by SF in the lower wells of microwell modified Boyden chambers. Assays were performed as described by Rosen et al. (1990b). Values plotted represent the numbers of cells migrating from the upper wells to the underside of the filter during a 6-hr incubation (mean SEM of triplicate assays). (B) Urokinase production. Confluent BBECs were incubated with SF for 24 hr. The medium was replaced with fresh DMEM (0.lml/cm2), and the cells were incubated for 6 hr. The 6-hr-conditioned medium was collected, and the cells were lysed. Plasminogen activator activity was assayed as described by Grant et al. (1993), and the total activity (secreted plus cell-associated) was normalized per microgram of cell protein. Nearly all of this activity was blocked by antiurokinase antibodies, indicating that it is due to uPA rather than tPA. (C) Invasion. BBECs were induced to migrate invasion through porous filters coated with a

*

SCATTER FACTOR AND ANGIOGENESIS

263

termined by reverse transcriptase PCR analysis of pericyte RNA (E. M. Rosen and M. Park, unpublished results), which is consistent with the finding that pericytes express immunoreactive c-met protein in vivo. Moreover, we have found that SF stimulates the proliferation of bovine aortic SMCs and bovine retinal pericytes in vitro (Fig. 2). The maximum degrees of stimulation of proliferation in media containing 1 and 5% calf serum were 2.4- and 1.8-fold for SMCs, 1.7- and 2.2-fold for pericytes, and 1.7- and 1.5-fold for bovine capillary ECs, respectively. These maximal values were observed at 20-100 ng/ml SF. These findings are consistent with the putative role of SMCs in angiogenesis and the presence of SMCs in new microvessels induced by SF (Section 1I.C). T h e ability of SMCs to produce and respond to SF suggests that SF may function as an autocrine growth factor for this cell type. However, the phenotype of cultured SMCs does not necessarily reflect that of SMCs present in normal adult large blood vessels. Normal adult SMCs in vivo are usually quiescent, but can be stimulated to proliferate by vascular injury. A parallel phenomenon called phenotypic modulation occurs when SMCs are explanted from vessels and passages in vitro (Campbell and Chamley-Campbell, 1981). Phenotypic modulation is characterized by transition from a quiescent contractile phenotype to an active synthetic-proliferative phenotype. Thus, cultured SMCs may be characteristic of proliferating injured SMCs rather than nonproliferating contractile SMCs. It seems unlikely that normal, unstimulated SMCs produce significant amounts of SF in vivo. Our findings suggest that SF may mediate paracrine (SMC-EC) and autocrine (SMC-SMC) interactions associated with the vascular response to injury. C. INVIVOANCIOGENIC ACTIVITY We used two different assays, the mouse Matrigel assay and the rat cornea assay, to demonstrate the ability of SF to induce new blood vessel formation in vivo (Grant et al., 1993; Naidu et al., 1994). In the former,

basement membrane matrix (Matrigel)by SF in the lower wells of 0.2-1311blind well Boyden chambers. Assays were performed as described by Bhargava et al. (1992). Values plotted represent the numbers of cells migrating to the lower surfaces of the filters during a 48-hr incubation (mean ? SEM of triplicate assays).(D) Capillary-like tube formation. HUVECs were induced to form tubes on Matrigel-coatedsurfaces by SF. Assays were performed as described by Rosen et al. (1991b). Values plotted represent the total tube length after a 24hr incubation, as determined by computerized digital imaging (mean ? range of duplicate assays).

264

ELIOT M. ROSEN AND ITZHAK D. GOLDBERG

100

F 1 -

r

-

-0- 5%CS

?5 IY

50

fz

40

w m -I -I

g

20

0

SCATTER FACTOR CONCENTRATION (ng/rnl)

FIG. 2. Effect of SF on proliferation of vascular cell types. Bovine aortic smooth muscle cells (SMCs) (A), bovine retinal pericytes (B), or bovine adrenal capillary endothelial cells (BCEs) (C) were seeded into 2-cm2 culture wells at 1 x lo4 (SMCs, BCEs) or 2 x 104 (pericytes) cells per well in DMEM containing 10%calf serum. The cells were allowed to attach overnight, after which the medium was replaced with DMEM containing 1 or 5% calf serum and the indicated dose of SF. Cells were incubated for 3 (SMCs, BCEs) or 6 days (with one refeeding on day 3) for slower growing pericytes. Values represent cell counts from triplicate wells (mean ? SD).

different doses of SF were mixed with 0.5 ml of Matrigel in the liquid state at 4°C. T h e Matrigel was injected subcutaneously into either XID nude beige mice or C57/BL mice. At body temperature, Matrigel rapidly forms a solid gel, retaining the SF and allowing prolonged exposure of the surrounding tissues to it. Animals were sacrificed after 10 days, and the ingrowth of blood vessels into the Matrigel plugs was quantitated by computerized digital image analysis of histological sections stained with Masson’s trichrome. Angiogenesis assessed at day 10 increased in a dosedependent manner from 2-200 ng/ml SF to 4-5 times control values. Responses were quantitatively similar in nude mice and C57/BL mice. Inflammatory responses were not observed in nude mice at any SF dose and were found only at supramaximal SFdoses (22000 ng/nl) in C57/BL mice. In the second assay, SF was dissolved in Hydron polymer, and dried Hydron pellets were placed in surgically created pockets about 1.5 mm from the limbus of the avascular rat cornea. Animals were perfused with colloidal carbon and sacrificed after 7 days. The growth of new vessels from the limbus toward the pellet was assessed. In these assays, SF in-

SCATTER FACTOR AND ANGIOGENESIS

265

duced dose-dependent corneal neovascularization in a fashion similar to that observed in the murine assay (Fig. 3). Purified native mouse SF and recombinant human SF induced equal angiogeneic responses, and the maximal responses induced by SF were similar in intensity to that induced by basic FGF (Grant et al., 1993). Antibodies against SF blocked SF-induced angiogenesis but did not affect FGF-induced angiogenesis.

FIG. 3. Neovascular responses induced in the rat cornea by purified mouse SF. Hydron pellets containing test samples were implanted into rat corneas to allow ingrowth of vessels from the limbus. After 7 days, corneas were perfused with colloidal carbon, and whole mount preparations were photographed. The darkened central area is the reflection of the implant. No angiogenic responses were observed in control pellets containing PBS (A), while at 50 ng of SF, the response was weak but positive (B). Strong positive responses were seen at 100 and 500 ng of SF (C and D), which are comparable to that induced by 150 ng of basic FGF (E), a positive control. These responses consist of sustained directional ingrowth of capillary sprouts and hairpin loops surrounding the implants. A, Control; B, Scatter factor, 50 ng; C, Scatter factor, 100 ng; D, Scatter factor, 500 ng; E, Basic FGF, 150 ng.

266

ELIOT M . ROSEN AND ITZHAK D. GOLDBERG

Inflammatory responses, assessed by F4/80 immunostaining to detect monocyte-macrophage infiltration, were observed only at supramaxima1 doses of SF. Therefore, SF appears to be at least as potent an inducer of angiogenesis as basic FGF. Our findings indicate that the recruitment of inflammatory cells does not play a major role in SF-induced angiogenesis. However, in the mouse Matrigel assay, histologic sections prepared at early times (days 2-3) revealed many SMC/pericyte-like cells present in the Matrigel. Moreover, at higher doses of SF, histologic sections prepared on day 10 revealed SMCs in some of the newly formed vessels in the Matrigel (Grant et al., 1993). Thus, SF-induced angiogenesis appears to be mediated by direct effects on ECs and, in addition, by direct and/or indirect effects on SMCs. Ill. SF as a Potential Tumor Angiogenesis Factor

A. ANCIOCENESIS IN HUMAN CANCERS Recent clinical studies suggest that tumor angiogenesis, as indicated by increased numbers of microvessels in the tumor stroma, is a strong independent indicator of poor prognosis in patients with invasive breast cancer (Weidner et al., 1991b, 1992; Bosari et al., 1992; Toi et al., 1993). A subset of patients with noninvasive breast tumors [ductal carcinoma znsztu (DCIS)] also exhibit elevated microvessel counts. Increased vessel count in DCIS patients is highly associated with other features suggestive of aggressive tumor biology (e.g., comedo subtype of DCIS, HERPlneu oncoprotein expression, and high KiSl proliferation index) (Guidi et al., 1994). Experimental studies of human and animal tissues indicate that an angiogenic phenotype may be observed in even earlier lesions (e.g., hyperplasia o r dysplasia) of breast or other tissues (Brem et al., 1978; Folkman et al., 1989). It has not been established whether angiogenesis is required for early progression from the noninvasive to invasive cancer phenotype or whether it merely reflects an underlying aggressive tumor biology. However, various studies suggest that angiogenesis is a critical requirement for local growth and metastasis of established solid tumors ( Folkman, 1992). Whereas physiological angiogenesis in normal adult tissues (e.g., as occurs during wound healing, corpus luteum formation, and placental implantation) is tightly regulated spatially and temporally, tumor angiogenesis is characterized by persistent, abnormal neovascularization. A modest number of growth factors and cytokines are capable of inducing

SCATTER FACTOR AND ANGIOGENESIS

267

angiogenesis in various in uiuo and in uitro assay systems [e.g., angiogenin, FGFs, EGF/TGFa, IL-8, PDECGF (platelet-derived endothelial cell growth factor), SF, TNFa, TGFP, and VEGF]. These angiogenic factors may be produced by tumor cells, host stromal cells (e.g., fibroblasts and SMCs), o r infiltrating leukocytes (e.g., lymphocytes, macrophages, and mast cells) (Polverini, 1989; Leek et al., 1994). The precise mechanisms leading to angiogenesis in human cancers, the specific angiogenic factor(s) involved, and the cell types that produce them are not welldelineated. OF SF WITHIN TUMORS B. EXPRESSION

1 . Carcinomas Both SE and c-met appear to be up-regulated and down-regulated in precisely coordinated patterns during normal developmental and reparative processes (Sonnenberg et al., 1993; Matsumoto and Nakamura, 1993;Joannidis et al., 1994). On the other hand, SF is chronically overexpressed in tumors (Rosen et al., 1994b; Yamashita et al., 1994; Joseph et al., 1995). As described in the following, overproduction of SF within tumors may be due, in part, to the accumulation of specific SF-inducing proteins. A high titer of SF in extracts of primary invasive breast carcinomas was found to be a powerful independent predictor of relapse and death (Yamashita et al., 1994). In patients with transitional cell bladder cancers, higher titers of SF were found in high grade, muscle-invasive cancers than in low grade noninvasive or superficially invasive cancers (Joseph et al., 1995). Patients in the former category usually fared poorly in comparison with patients in the latter category. Since both SF content and tumor angiogenesis are strong independent prognostic indicators for breast carcinoma, it is reasonable to speculate that SF may be functioning as a breast cancer angiogen. If, indeed, SF functions as a tumor angiogen, then future studies should reveal a close correlation between tumoral SF content and the quantitative extent of tumor angiogenesis. Such a correlation may not be exact since, as described earlier, various other factors may contribute to tumor angiogenesis. Moreover, several naturally occurring protein factors, including thrombospondin (TSPl) and platelet factor-4, function as inhibitors of angiogenesis (see the following). It is also likely that SF may interact additively or synergistically with other angiogenesis-inducing factors, such as VEGF. Since many or most of these factors are found in tumors, the net angiogenic phenotype of the tumor may be determined by the balance of proangiogenic and antiangiogenic factors present.

268

ELIOT M. ROSEN A N D ITZHAK D . GOLDBERG

2. Kaposi’s Sarcoma (KS)

KS is a multicellular neoplasm characterized by a major component of EC proliferation and angiogenesis. T h e epidemic form of KS occurs frequently in homosexual males with acquired immunodeficiency syndrome (AIDS). AIDS-KS is usually regarded as a multifocal cytokinedependent tumor rather than a malignant metastasizing cancer (Mallery et al., 1994). In a recent study, SF was found to mediate several functions that may be related to the pathogenesis of AIDS-related KS: (1) in vitro transdifferentiation of normal human ECs into a KS tumor cell-like phenotype and (2) autocrine stimulation of KS tumor cell proliferation. In addition, SF was found to be the major angiogenic molecule present in conditioned medium from human T-lymphotrophic virus type IIinfected T lymphocytes (HTLV-I1 CM) (Naidu et al., 1994). HTLV-I1 C M is required for the long-term cultivation of KS tumor cells. These findings suggest that SF may contribute to EC proliferation and angiogenesis in KS tumors. The potential role of SF in KS is discussed more fully in a prior volume (Polverini and Nickoloff, 1995). C. MECHANISMS OF TUMORAL SF PRODUCTION 1 . Cellular Sources of Tumor SF

In rritro, the major SF-producing cell types are mesenchymal cells, including fibroblasts, vascular smooth muscle cells, glial cells, macrophages, endothelial cells, and T lymphocytes (Stoker et al., 1987; Rosen et al., 1989, 1994a; Noji et al., 1990; Shiota et al., 1992; Yanagita et al., 1992; Naidu et al., 1994). Many of these cell types are present within tumor stroma and so may contribute to the accumulation of SF in solid tumors. Epithelial cells generally are nonproducers, although modest titers of SF are produced by some epithelial and carcinoma cell lines (Adams et al., 1991; Tsao et al., 1993).

FIG.4. Expression of SF and c-met proteins in high grade, muscle-invasive human transitional cell carcinomas of the bladder. Paraffin-embedded sections were immunoperoxidase-stained using a Vectastain ABC kit (Bector Labs, Burlington, VT) and counterstained with hematoxylin. T h e primary antibodies were sheep antiserum to human SF ( 1 :500) (A-C;), rabbit antiserum to a c-met C-terminal peptide ( 1 :500) (D and E), and rabbit nonimmune serum ( 1 :500) (F). Strong SF staining can be seen in carcinoma cells (A-C), blood vessel wall cells (A and B), and bladder wall smooth muscle cells (C).Strong c-met staining can be found in carcinoma cells (D and E), blood vessel wall cells (E), and bladder wall smooth muscle (D and E). Nonimmune controls showed no staining (F). Scale: 100 km = 0.32 cm (A), 1.27 cm (B-E), 0.64 crn (F).

T

MARCOA. PIEROTTI, Daimion of Expenmental Oncology A , Istituto Nazaonale Tumorz, 20133 Malan, Italy (83) ELioT M. ROSEN, Department of Radaation Oncology, Long Island Jewzsh Medacal Center, The Long Island Campiu for Albert Eanstezn College of Medicane, “Vew Hyde Park, New York 11042 (257) FRANCISCORUIZ-CABELLO, S e n i m o de Ancilsu Clinacos e Inmunologia, Hospital Virgen de las Nzezies, Unzuerszdad de Granada, 18014 Granada, Spuan (155) MASABUMISHIBUYA, Institute of Medical Science, Unzversity of Tokyo, Tokyo 108, Japan (281)

This Page Intentionally Left Blank

269

SCATTER FACTOR AND ANGIOGENESIS

We have observed positive immunostaining of carcinoma cells, bladder wall smooth muscle, fibroblasts, and vascular wall cells (SMCs and ECs) for SF in transitional cell carcinomas of bladder (see Fig. 4A-4C). However, as is true for most carcinoma cell types, transitional carcinoma cells do not produce any SF in uitro (Joseph et al., 1995). Since SF is a soluble cytokine, the in uiuo staining of carcinoma cells may result from uptake of the factor rather than direct synthesis. Alternatively, the tumor cells may lose the ability to synthesize SF during adaptation to cell culture. 2. Role of Cellular Activation As described in Section II.A.l, the synthesis of SF by endothelium in local and distant capillary beds is up-regulated following injury to the liver. T h e mechanisms responsible have not been identified, but may include exposure to some of the cytokines and soluble factors described here. Purified resting human T lymphocytes and the HUT 78 human T-cell line produce little or no SF. In contrast, the retrovirus (HTLV-11)infected human T-cell line 38-10 expresses SF mRNA and secretes large quantities of bioactive and immunoreactive SF (Naidu et al., 1994). Retroviral infection is known to activate T cells and to induce expression by these cells of a variety of cytokines (Fulton et al., 1987). Tumor-associated macrophages are thought to contribute to tumor angiogenesis by secretion of angiogenic cytokines such as TNFa (Pol-

PRODUCTION OF SF

BY

TABLE I MONOCYTEAND MACROPHAGE CELLPOPULATIONS~

Cell type and source Monocytic cells Human peripheral blood monocytes T H P l human monocytic leukemia line U937 human monocytic cell line Macrophages Resident peritoneal macrophages from Balb/c mice Rat resident peritoneal macrophages Human macrophages (after in vztro differentiation of monocytes) Bac2F5 mouse macrophage cell line

SF production rate (units/l06 cells/48 hr)

<2 2-4 2-4 8-16 8-16 16 16-32

a SF was assayed using the MDCK serial dilution scatter assay, a sensitive and specific bioassay (Rosen et al., 1990a,b). The SF titer at the limiting dilution of the assay was defined as 0.5 MDCK scatter units per millilter.

270

ELIOT M. ROSEN AND ITZHAK D. GOLDBERG

verini, 1989). Macrophages and macrophage-like cells express SF mRNA and produce immunoreactive and biologically active SF (Noji et at., 1990; Wolf et al., 1991; Yanagita et al., 1992; Inaba at al., 1993; Rosen et al., 1994a). Our findings suggest that monocytes normally produce little SF, but acquire the ability to produce SF when they undergo conversion to macrophages (Table I). Since the level of SF production may be related to the degree of cell activation, we expect that tumor macrophages, which are highly activated, would exhibit even higher levels of SF production than normal tissue-derived (e.g., peritoneal) macrophages. 3 . Regulation of SF Production by Cytokines and SF-Inducing Factors (SF-IFs) SF production by some normal human fibroblast types (e.g., lung and gingival fibroblasts) is stimulated by the proinflammatory cytokines IL-1 and TNFa (Tamura et al., 1993) and inhibited by the antiinflammatory cytokine TGFP (Gohda et al., 1992). These cytokines may be found within tumors and so may influence tumoral SF accumulation. However, studies suggest that other SF regulatory proteins are present in the

SF-IFs Autocrine

LAM ____o

Stromal

@

Scatter Factor

T N T ; Fibroblast

bb

Tumor Angiogenesis

SF-IFs

Cells

CSF-I

Chemoattractant

?SF

- - - -b

I

u

.-...- .

~

Scatter Factor

M~~~ .--.-phage

Tumor Cells Stromal Cells Endothelial Cells

p a c ro p h a g e s

I

I

----0

Altered Expression of Angiogene.sis Regulatory Molecules (eg., TSPI)

FIG. 5. Putatibre mechanisms of SF production in tumors. Carcinoma cells secrete factors (SF-IFs) that induce fibroblasts to produce SF. Fibroblasts also secrete their own

autocrine SF-IFs. Carcinoma cells produce chemoattractants for monocytes (Ramakrishnan t t al.. I YW), which accumulate in the tumor as macrophages. Macrophages produce SF and proinflammatory cytokines that further stimulate SF production. SF may also induce alterations in the synthesis of other angiogenesis regulatory factors by various cell types in the t u m o r .

SCATTER FACTOR AND ANGIOGENESIS

271

tumor environment. Several studies indicate that mouse and human mammary carcinoma cells, which do not produce SF, produce soluble factors distinct from IL-1 and TNFa that stimulate SF mRNA aqd protein expression by fibroblasts and other stromal cell types (Seslar et al., 1993; Rosen et al., 1994a,b).At least two distinct factors are produced: a high-molecular-weight (>30 kDa) heat-labile protein activity and a lowmolecular-weight (<30 kDa) heat-stable protein activity. A 10-30-kDa heat-stable SF-inducing activity called injurin appears in the serum of rats within several hours of liver injury (Matsumoto et al., 1992). Injurin may be homologous to the <3O-kDa tumor-produced SF-IF protein(s), but its purification and further characterization have not been reported. We have definitively purified and characterized a new and unique 12kDa SF-inducing protein from a high producer clone of ras oncogenetransformed NIH2/3T3 mouse cells (Rosen et al., 1994a). This ras-3T3 SF-IF protein has properties similar to those of the <30-kDa factor secreted by breast carcinoma cells and injurin. It acts at physiological concentrations (20-400 pM) to stimulate SF mRNA and protein expression by up to 4-6-fold. While breast carcinoma cells produce both high and low-molecular-weightSF-IFs (Seslar et al., 1993; Rosen et al., 1994a,b), transitional cell bladder carcinomas produce a >30-kDa SF-IF protein(s) nearly exclusively (Joseph et al., 1995). A similar >3O-kDa SF-IF is present in extracts from bladder cancers and in urine from bladder cancer patients. On the other hand, little or no SF-inducingactivity is present in urine from control patients. On the basis of these findings, we hypothesize that SF is overproduced in tumors via an abnormal tumor-stroma interaction in which tumor cells secrete factors (SF-IFs) that induce stroma1 cell SF production (Fig. 5). IV. Role of SF in Angiogenesis: Hypotheses and Future Directions

A. LINKAGE OF ANGIOCENESIS AND TUMOR SUPPRESSORS 1 . Thrombospondin ( T S P l )

TSPl is a complex, multidomain adhesive glycoprotein of the extracellular matrix that mediates cell-cell and cell-matrix interactions (Lawler, 1993).A portion of the TSPl molecule with angioinhibitory activity (GP140) is encoded by a putative tumor suppressor gene that is inactivated when BHK hamster cells are chemically transformed into a tumorigenic phenotype (Rastinejad et al., 1989; Good et al., 1990). GP140 and native TSP 1 inhibit EC proliferation, EC migration, and angiogenesis

272

ELIOT M. ROSEN AND ITZHAK D. GOLDBERG

induced by basic FGF, TNFa, and macrophage conditioned medium (Good et al., 1990; Taraboletti et al., 1990). Preliminary findings from our laboratory indicate that TSPl also blocks SF-induced migration of capillary ECs and SF-induced neovascularization in the rat cornea (P. J. Polverini and E. M. Rosen, unpublished results). The mechanism of TSPl's antiangiogenic activity has not been fully defined. However, a study suggests that all of this activity is concentrated in several 15-20-amino acid regions within the procollagen homology domain and type I repeats of the central stalk of the molecule (Tolsma et al., 1993). Thus, a critical mutation corresponding to one of these sites might render TSP 1 unable to block angiogenesis. TSPl accumulates in human tumors, such as breast cancers (Wong et al., 1992; Clezardin et al., 1993). However, its function in these tumors is unknown, nor is it known whether function-negative TSPl mutants are present in some of these tumors. 2 . p53 Tumor Suppressor Gene T h e p53 antioncogene encodes a nuclear phosphoprotein that generally regulates cell growth by transactivating o r repressing transcription (Tominaga et al., 1992; Mack et ad., 1993). Loss of p53 function in LiFraumeni cancer syndrome fibroblasts results in decreased expression of TSPl and acquisition of an angiogenic phenotype, whereas transfection of a wild-type p53 expression construct into fibroblasts lacking a functional p53 allele results in up-regulation of TSPl and loss of the angiogenic phenotype (Dameron et al., 1994). These findings suggest that tumor suppressor gene products can directly o r indirectly regulate the angiogenic phenotype. p53 inhibits transcription of the bcl-2 protooncogene, whose protein product suppresses apoptosis (programmed cell death). p53 also induces the transcription of bax, a gene whose protein product binds to bcl-2 protein and inhibits its biologic activity (Miyashita et al., 1994). Both bcl-2 overexpression and SF can overcome apoptosis associated with detachment of epithelial cells from their substrata (Frisch and Francis, 1994). Detachment of epithelial and vascular endothelial cells from the underlying basement membrane is an early step in tumor cell invasion and in angiogenesis, respectively. In Fig. 6, we propose several mechanisms by which the loss of tumor suppressor gene function may promote SFmediated invasion and angiogenesis in tumors. B. SF-

AND C-met-INDUCING

FACTORS

If SF is an important angiogenic molecule in tumors, the role of SFIFs and the rnechanism(s) by which SF-IF synthesis is regulated become

273

SCATTER FACTOR AND ANGIOGENESIS

critical issues. In Fig. 6, we postulated that the production of SF-IF(s) by tumor cells is activated by a tumor suppressor gene mutation(s) (e.g., p53), which leads to the loss of normal transcriptional repression. Alternatively, o r in addition, SF-IFs may be induced by mutations leading to the activation of oncogenes or by soluble factors (?IL-1 or TNFa) present in the tumor. In a study, c-met mRNA in ovarian carcinoma cells was found to have a very short half-life (30 min), and c-met mRNA expression was stimulated by a variety of soluble proinflammatory cytokines (IL-la, IL-6, and TNFa), as well as steroid hormones (Moghul et al., 1994). The existence of soluble inducers of both SF and c-met is consis-

I

NORMAL EPITHELIUM OR NON-INVASIVETUMOR P53 Mutation Other Tumor Suppressor Gene Mutations

ALTERED REGULATION OF GENE TRANSCRIPTION

?

Stromal Cell SF Production

? Expression of SF-IFs

? Endothelial Response to SF

I

Loss of Epithelial Requirement

& Angio-Inhibitory Activity and/or Other Putative Induced Angiogenesis

FIG. 6. Possible mechanismsthrough which loss of tumor supressor gene function may lead to SF-mediated invasion and angiogenesis in carcinomas.

274

ELIOT M. ROSEN AND ITZHAK D. GOLDBERG

tent with evidence suggesting that these genes are coregulated (Sonnenberg %et al., 1993;Joannidis et al., 1994).The existence of a soluble c-met inducer may also explain why tumor stromal cells (e.g., bladder wall SMCs, vascular SMCs, and vascular ECs) express high levels of immunoreactive c-met protein (see Figs. 4D,4E). Overexpression of c-met in these cell types may contribute to persistent angiogenesis and stromal growth by allowing these cell types to continue to respond vigorously to SF. C. DIRECTIONS FOR FURTHER EXPERIMENTAL STUDY

On the basis of the foregoing discussion, we postulate that (1) mutations of tumor suppressor genes (e.g., p53) induce activation of the SFIF+SF-+c-met paracrine loop; (2) the resulting overexpression of SF contributes to persistent neovascularization observed in tumors; and (3) expression of angioinhibitory molecules (e.g., TSP-1) in this setting is insufficient to neutralize the proangiogenic activity of SF and other tumor angiogens in biologically aggressive tumors. To further delineate the contribution of SF to tumor angiogenesis, we propose the following as potentially important areas for future study: (1) the mechanisms by which SF and c-met expression are regulated in tumors; (2) the linkage of these regulatory mechanisms to tumor suppressor systems; and (3) interactions between SF and tumor suppressor gene-encoded angiogenesis inhibitors (e.g., TSP1). D. POTENTIAL CLINICAL APPLICATIONS

If SF were found to be a clinically significant human angiogenesis factor, two types of clinical applications could be envisioned. First, exogenous SF may be supplied in situations where additional angiogenesis would be beneficial (e.g., to promote healing of burns, viability of transplanted tissues or organs, and regeneration of ischemic myocardium). Second, inhibition of SF activity might be used in the treatment of angiogenesis-dependent diseases. The latter include certain chronic inflammatory disorders (e.g., rheumatoid arthritis) and most malignant solid tumors (Polverini, 1989; Folkman, 1992). A truncated form of SF consisting of only the N-terminal loop and the first two kringle domains (NK2) is naturally produced by some human fibroblast types (Chan et al., 1991). NK2 is encoded by a 1.5-kb mRNA generated from the major 6-kb SF transcript by alternative mRN A processing. N K2 exhibits high-affinity binding to cellular c-met receptors and blocks SF mitogenic activity for epithelial cells. An even

SCATTER FACTOR AND ANGIOGENESIS

275

smaller recombinantly produced fragment of SF (NK1) exhibits similar properties (Lokker et al., 1992). If these molecules also cause endothelial cell SF receptor blockade, they may be useful antiangiogenesis agents. Finally, a recombinant chimeric antibody has been produced that consists of bivalent extracellular binding domains of the c-met receptor fused to the constant region of the human IgG heavy chain (Mark et al., 1992). This antibody neutralizes the binding of SF to c-met receptor on epithelial cells. It has three properties that may be important for clinical use as an anti-SF agent: (1) high-affinity interaction with the receptor-binding domain of SF; (2) production as a single purified antibody protein rather than a crude mixture; and (3) production of large quantities of antibody by recombinant DNA technology. When recombinant SF-IF and SF-IF blockers are available, they might be used as primary or supplementary agents to induce or inhibit endogenous SF production in order to stimulate or inhibit angiogenesis, respectively. These agents might also be combined with SF cleavage enzyme (HGF activator) or cleavage enzyme blockers to enhance or inhibit the conversion of endogenous inactive single-chain SF to biologically active two-chain SF. Thus, a variety of different approaches might be utilized to promote or inhibit SF-mediated angiogenesis in various clinical conditions.

V. Summary and Conclusions Scatter factor (hepatocyte growth factor) is a mesenchyme-derived cytokine that stimulates motility, proliferation, and morphogenesis of epithelia. These responses are transduced through the c-met protooncogene product, a transmembrane tyrosine kinase that functions as the SF receptor. SF is a potent angiogenic molecule, and its angiogenic activity is mediated primarily through direct actions on endothelial cells. These include stimulation of cell motility, proliferation, protease production, invasion, and organization into capillary-like tubes. SF is chronically overexpressed in tumors, suggesting that it may function as a tumor angiogenesis factor. SF production in tumors may be due, in part, to an abnormal tumor-stroma interaction, in which the tumor cells secrete factors (SF-IFs)that stimulate SF production by tumor-associated stromal cells. Studies suggest a link between tumor suppressors (antioncogenes) and inhibition of angiogenesis. We hypothesize that tumor suppressor gene mutations may contribute to the activation of an SFIF+SF+c-met pathway, leading to an invasive and angiogenic tumor phenotype. Modulation of this pathway may, ultimately, provide clinically useful methods of enhancing or inhibiting angiogenesis.

276

ELIOT M. ROSEN AND ITZHAK D. GOLDBERG

ACKNOWLEDGMENTS This work was supported in part by the USPHS (CA50516 and CA64869). Dr. Rosen is an Established Investigator of the American Heart Association (AHA 90-195). We thank Genentech, Inc., South San Francisco, CA, for providing recombinant human SF and sheep antiserum to human SF. We thank Dr. George Vande Woude, NCI-Frederick, Frederick, MD, for providing rabbit antiserum to human c-met peptide. We are grateful to Dr. James Kinsella, National Institute o n Aging, Baltimore, MD, for performing the capillary tube formation assays. We thank Drs. Patricia DAmore and Andrea Dodge, Department of Surgery, Children's Hospital, Boston, MA, for performing proliferation assays of smooth muscle cells, pericytes, a n d capillary endothelium a n d for providing pericyte cultures. PCR analysis of pericyte cDNA was performed by Dr. Morag Park and Monica Naujokas, Royal Victoria Hospital and McGill University, Montreal, Quebec.

REFERENCES

Adanis, J. C., Furlong, R. A., and Watt, F. M. (1991).J. Cell. Sci. 98, 385-394. hntonelli-Orlidge, A,, Smith, S. R., and DAmore, P. A. (1989). Am. Rev. Resp. Dis. 140, 1129-1131. Bhargava, M., Joseph, A., Knesel, J., Halaban, R., Li, Y., Pang, S., Goldberg, I., Setter, E., Donovan, M. A., Zarnegar, R., Michalopoulos, G . A., Nakamura, T., Faletto, D., and Rosen, E. M. (1992). Cell Growth Dtf,,eren. 3, 11-20. Bosari, S., Lee, A. K., DeLeilis, R. A., Wiley. B. D., Heatley, G . J., and Silverman, M. L. (1992). Hum. Pathol. 23, 755-761. Bottaro, D. P., Rubin, J. S., Faletto, D. L.. Chan, A. M.-L., Kmiecik, 'I. E., vdnde Woude, G. F., and Aaronsen, S. A. (1991). Scknce 251, 802-804. Brern, S. S., Jensen, H. M., and Gullino, P. M. (1978). Cancer 41, 239-244. Bussolino, F., DiRenzo, M. F., Ziche, M., Bocchieto, E., Olivero, M., Naldini, L., Gaudino, G., Tamagnone, L., Coffer, A., and Comoglio, P. M. (1992).J Cell. Biol. 119, 629-641. Campbell, G. R., and Chamley-Campbell, J. H . (1981). Med. Hypothses 7 , 729-735. Chan, A. M.-L., Rubin, J. S., Bottaro, D. P., Hirschfield, D. W., Chedid, M., and Aaronson, S. A. (1991).Science 254, 1382-1385. Clezardin, P., Frappart, L., Clerget, M., Pechoux, P., and Delnias, P. D. (1993). CancerRes. 53, 1421- 1430. Dameron, K. M., Volpert, 0 . V., Tainsky, M. A,, and Bouck, N. (1994). Science 265, 15821590. Folkman, J. (1985). Adv. Cancer Res. 43, 175-203. Folkman, J. (1992). Semin. Cancer B i d . 3, 65-71. Folkman, J., Watson, K., Ingber, D., and Hanahan, D. (1989). Nature 339, 58-61. Frisch, S. M.,a n d Francis, H . (1994).J. Cell. Biol. 144, 619-626. Fulton, R.. Forrest, D.. McFarlane, R., Onions, D., and Neil, J. C . (1987). Nature 326, 190194. Cherardi, E., Gray, J.. Stoker, M., Perryman. M., a n d Furlong, R. (1989). Proc. Natl. Acad. Sci. US4 86, 5844-5848. Gohda, E., Matsunaga, T., Kataoka, H., and Yamamoto, I. (1992). Cell. Biol. Int. Rep. 16, 9 17-926. Gonzatti-Haces, M., Seth, A., Park, M., Copeland, T., Oroszlan, S., and Vande Woude, G . F. (1988). Pror. Natl. Acad. Sci. USA 85, 21-25.

SCATTER FACTOR AND ANGIOGENESIS

277

Good, D. J., Polverini, P. J., Rastinejad, F., Le Beau, M. M., Lemons, R. S., Frazier, W. A,, and Bouck, N. P. (1990). Proc. Natl. Acad. Sci. USA 87, 6624-6628. Grant, D. S., Tashiro, K. I., Segui-Real, B., Yamada, Y., Martin, G. R., and Kleinman, H. K. (1989). Cell 58, 933-943. Grant, D. S., Kleinman, H. K., Goldberg, I. D., Bhargava, M., Nickoloff, B. J., Polverini, P., and Rosen, E. M. (1 993). Proc. Natl. Acad. Sn‘. USA 90, 1937- 194 1 . Guidi, A. J., Fischer, L., Harris, J. R., and Schnitt, S. J. (1994).JNCI 86, 356-361. Halaban, R., Rubin, J., Funasaka, Y., Cobb, M., Boulton, T., Faletto, D., Rosen, E., Chan, A., Yoko, K., White, W., Cook, C., and Moellmann, G. (1992). Oncogene 7, 21952206. Huff, J. L., Jelinek, M. A., Borgman, C. A., and Lansing, T. J. (1994).Proc. Natl. Acad. Sci. USA 90,6140-6144. Inaba, M., Koyama, H., Hino, M., Okuno, S., Terada, M., Nishizawa, Y., Nishino, T., and Morii, H. (1993). Blood 82, 53-59. Joannidis, M., Spokes, K., Nakamura, T., Faletto, D., and Cantley, L. G. (1994). Am. J. Physiol. 267, F231-F236. Joseph, A., Weiss, G. H., Jin, L., Fuchs, A., Chowdhury, S., OShaughnessy, P., Goldberg, I. D., and Rosen, E. M. (1995).J. Natl. Cancer Inst. 87, 372-377. Kan, M., Zhang, G. H., Zarnegar, R., Michalapoulos, G., Myoken, Y., McKeehan, W. L., and Stevens, J. L. (1991). Biochem. Biqhys. Res. Commun. 174, 331-337. Lawler, J. (1993). I n “Guidebook to the Extracellular Matrix and Adhesion Proteins” (T. Kreis and R. Vale, Eds.), pp. 95-96. Oxford University Press, Inc., New York. Leek, R. D., Harris, A. L., and Lewis, C. E. (1994).J. L e d . B i d . 56,423-435. Li, Y., Bhargava, M. M., Joseph, A., Jin, L., Rosen, E. M., and Goldberg, I. D. (1994). I n Vitro Cell Dev. B i d . 30A, 105-1 10. Lokker, N. A., Mark, M. R., Luis, E. A., Bennett, G. L., Robbins, K. A., Baker, J. B., and Godowski, P. J. (1992). E M B O J . 11, 2403-2410. Mack, D. H., Vartikar, J., Pipas, J. M., and Laimins, L. A. (1993). Nature 363, 281-283. Mallery, S. R., Lantry, L. E., Hegtvedt, A., Lazo, A., Titterington, L. C., Hout, B. L., Brierley, G. P., and Stephens, R. E. (1994). Lymphology 27, 45-48. Mark, M. R., Lokker, N. A., Zioncheck, T. F., Luis, E. A., andGodowski, P. J. (1992).J.B i d . Chem. 267, 26166-26171. Mars, W. M., Zarnegar, R., and Michalopoulos, G. K. (1993).A m . ] . Puthol. 145,949-958. Matsumoto, K., and Nakamura, T. (1993). I n “Hepatocyte Growth Factor-Scatter Factor and the c-Met Receptor” (I. D. Goldberg and E. M. Rosen, Eds.), pp, 225-250. Birkhauser-Verlag, Basel. Matsumoto, K., Tajima, H., Hamanoue, M., Kohno, S., Kinoshita, T., and Nakamura, T. (1992). Proc. Natl. Acad. Sci. USA 89, 3800-3804. Matsumoto, K., Tajima, H., Okazaki, R., and Nakamura, T. (1993).J.Biochem. (Tokyo) 114, 820-826. Miyashita, T., Krajewski, S., Krajewska, M., Wang, H. G., Lin, H. K., Liebermann, D. A., Hoffman, B., and Reed, J. C. (1994). Oncogene 9, 1799-1805. Miyazawa, K., Tsubouchi, H., Naka, D., Takahashi, K., Okigaki, M., Arakaki, N., Nakayama, S., Hirono, S., Sakiyama, O., Gohda, E., Daikuhara, y., and Kitamura, N. (1989). Biochem. Biophys. Res. Commun. 163, 967-973. Miyazawa, K., Shimomura, T., Kitamura, A,, Kondo, J., Morimoto, Y., and Kitamura, N. 1993).J. Biol. Chem. 268, 10024-10028. Miyazawa, K., Shimomura, T., Naka, D., and Kitamura, N. (1994).J.B i d . Chem. 269,89668970.

278

ELIOT M . ROSEN AND ITZHAK D. GOLDBERG

Moghul, A., Lin, L., Beedle, A., Kanbour-Shakir, A., DeFrances, M. C., Liu, Y., and Zarnegar, R. (1994). Oncogene 9, 2045-2052. Montesano, R., Matsurnoto, K., Nakarnura, T., and Orci, L. (1991). Cell 67, 901-908. Morimota, A., Okamura, K., Harnanaka, R., Sato, Y., Shima, N., Higashio, K., and Kuwano, M. (1991). Biochem. Bi@zys. Res. C m m u n . 179, 1042-1049. Naidu, Y. M., Rosen, E. M., Zitnik, R., Goldberg, I., Park, M., Naujokas, M., Polverini, P. J., and Nickoloff, B. J. (1994). Proc. Nutl. A d . Sri. USA 91, 5281-5285. Nakarnura, T., Nishizawa, T., Hagiya, M., Seki, T., Shimonishi, M., Sugimura, A,, and Shimizu, S. (1989). Nature 342, 440-443. Naldini, L., Tanlagnone, L., Vigna, E., Sachs, M.. Hartrnann, L.. Birchmeier, W., Daikuhara, Y., Tsubouchi, H., Blasi, F., and Cornoglio. P. M. (1992). E M B O J . 11, 48254833. Noji, S., Tashiro, K., Koyarna, E., Nohno, T., Ohyarna, K., Taniguchi, S., and Nakamura, T. (1990). Bioc-hem. Bzophys. Re5. Cummun. 173, 42-47. Pepper, M. S., Matsumoto, K., Nakamura, T., Orci, L., and Montesano, R. (1992).J. B i d . Chem. 267,20493-20496. Polverini, P. J, (1989). Cytokirws 1, 54-73. Polverini, P. J., and Nickoloff, B. J. (1995). Adv. Cancer Re5 66, in press. Ponzetto, C;., Bardelli, A., Zhen, Z., Maina, F., Zonca, P., Giordano, S., Graziani, A,, Panayoyou, G., and Cornoglio, P. M. (1994). Cell 77, 261-271. Rarnakrishnan, S., Xu, F. J.. Brandt, S. J., Niedel, J. E., Bast, R. C., Jr., and Brown, E. L. ( 1989).j . Clin. Invest. 83, 92 1-926. Rastinejad. F., Polverini, P. J., and Bouck, N. P. (1989). Cell 56, 345-3.55. Ronsin, C.. Muscatelli, F., Mattei, M.-G., and Breathnach, R. (1993). Oncogene 8, 11951202. Rosen, E. M., Goldberg, I. D., Kacinski, B. M., Buckholz, T., and Vinter, D. W. (1989). In Vitro C ~ lDev. l Biol. 25, 163- 173. Rosen, E. M., Merornsky, L., Setter, E., \’inter, D. W., and Coldberg, I. D. (1990a). Invasion M e t a \ t . 10, 49-64. Rosen, E. M., Meronisky. L., Setter, E., Vinter, D. W., and Goldberg, I. D. (1990b). Proc. Sor. Exp. Biol. Med. 195, 34-43. Exp. Cell. Rosen, E. M.. Merornsky, L., Setter, E., Vinter, D. W., and Goldberg, I. D. (1990~). Res. 186, 22-3 1. Rosen, E. M., Goldberg, I. D., Liu, D., Setter, E., Donovan, M. A,, Bhargava, M., Reiss, M., and Kacinski, B. M. (1991a). Cancer Res. 57, 5315-5321. Rosen. E. M., Grant, D., Kleinrnan, H., Jaken, S., Donovan, M. A., Setter, E., Luckett, P. M., and Carley, W. (1991b). In “Cell Motility Factors” (I. D. Goldberg and E. M. Rosen, Eds.), pp. 76-88. Birkhauser-Verlag, Basel. Rosen, E. M.,Jaken, S., Carley, W., Setter, E., Bhargava, M., and Goldberg, I. D. (lYYIc).]. Cell. Physiol. 146, 325-335. Rosen, E. M., Joseph, A., Jin. L., Rockwell, S., Elias, J. A., Knesel, J., Wines, J., McClellan, J.. Kluger, M.J., Goldberg, I. D., and Zitnik, R. (l994a).J. Cell. Bzol. 127, 225-234. Rosen, E. M.,Knesel, J.. Goldberg, I. D. Bhargava, M., Joseph, A., Zitnik, R., Wines, J., Kelley, M., and Rockwell, S. (1994b). Int.1. Cancer 57, 706-714. Kubin, J. S., Chan, A. M.-L., Bottaro, D. P., Burgess, W. H., Taylor, W. G., Cech, A. C., Hirschfield, D. W., Wong, J., Miki, T., Finch, P. W., and Aaronson, S. A. (1991). Proc. , V d . Acad. Sci. USA 88, 415-419. Saksela, 0.. and Rifkin, D. B. (1988). Annu. Rm. Cell. B i d . 4, 93-126. Santos, 0. F. P., Moira, L. A., Rosen, E. M., and Nigarn, S. K. (1993).Dev. B i d . 159, 538.543.

SCATTER FACTOR AND ANGIOGENESIS

279

Seslar, S. P., Nakamura, T., and Byers, S. W. (1993). Cancer Res. 53, 1233-1238. Shiota, G., Rhoads, D. B., Wang, T. C., Nakamura, T., and Schmidt E. V. (1992). Proc. Nad. Acad. Sci. USA 89, 373-377. Sonenberg, E., Meyer, D., Weidner, K. M., and Birchmeier, C. (1993).J. Cell. Biol. 123, 223-235. Stoker, M., Gherardi, E., Perryman, M., and Gray, J. (1987). Nature 327, 238-242. Tamura, M., Arakaki, N., Tsoubouchi, H., Takada, H., and Daikuhara, Y. (1993). J. Eiol. Chem. 268,8140-8145. Taraboletti, G., Roberts, D., Liotta, L. A., and Giavazzi, R. (1990). J. Cell. Biot. 111, 765772. Toi, M., Kashitani, J., and Tominaga, T. (1993). Int. J. Cancer 55, 371-374. Tolsma, S. S., Volpert, 0. V., Good, D. J., Frazier, W. A., Polverini, P. J., and Bouck, N. (1993).J. Cell. Biol. 122, 497-5 11. Tominaga, O., Hamelin, R., Remvikos, Y., Salmon, R.J., and Thomas, G. (1992). Crit. Rev. Oncogen. 3, 257-282. Tsao, M. S., Zhu, H., Giaid, A., Viallet,J., Nakamura, T., and Park, M. (1993). CellGiowth Differen. 4, 571-579. Tsarfaty, I., Resau, J. H., Rulong, S., Keydar, I., Faletto, D., and Vande Woude, G . F. (1992). Science 257, 1258-1261. Wang, M. H., Ronsin, C., Gesnel, M.-C., Coupey, L., Skeel, A., Leonard, E. J., and Breathnach, R. (1994). Science 266. 117-1 19. Weidner, K. M., Behrens, J., Vandekerckhove, J,, and Birchmeier, W. (1990).J. Cell. Biol. 111, 2097-2108. Weidner, K. M., Arakaki, N., Vandekereckhove,J., Weingart, S., Hartmann, G., Rieder, H., Fonatsch, C., Tsubouchi, H., Hishida, T., Daikuhara, Y., and Birchmeier, W. (1991a). Proc. Natl. Acad. Sci. USA 88, 7001-7005. Weidner, N., Semple,J. P.,Welch, W. R., and Folkman,J. (1991b). NauEngl.J. Med. 324, 1-8. Weidner, N., Folkman,J., Pozza, F., Bevilaqua, P., Allred, E. N., Moore, D. H., Meli, S., and Gasparini, G. (1992).J. Natl. Cancer Imt. 84, 1875-1887. Weidner, K. M., Sachs, M., and Birchmeier, W. (1993).J . Cell. Biol. 121, 145-154. Wolf, H. K., Zarnegar, R., and Michaelopoulos, G. K. (1991). Hepatology 14, 488-494. Wong, S. Y.,Purdie, A. T., and Han, P. (1992). Am. J. Pathol. 140, 1473-1482. Yamashita, J., Ogawa, M., Yamashita, S., Nomura, K., Kuramoto, M., Saishoji, T., and Sadahito, S. ( 1994). Cancer Res. 54, 1630-1 633. Yanagita, K., Nagaike, M., Ishibashi, H., Niho, Y., Matsumoto, K., and Nakamura, T. (1992). Biochem. Biophys. Res. Commun. 182, 802-809. Yoshimura, T., Yuhki, N., Wang, M. H., Skeel, A., and Leonard, E. J. (1993).J. Biol. Chem. 268, 15461-15468. Zhu, H., Naujokas, M. A., and Park, M. (1994). Cell Growth Differen. 5, 359-366.

This Page Intentionally Left Blank

ROLE OF VEGF-FLT RECEPTOR SYSTEM IN NORMAL AND TUMOR ANGIOGENESIS Masabumi Shibuya Institute of Medical Science, University of Tokyo, Tokyo 108, Japan

I. Introduction A. New System for the Regulation of Angiogenesis: VEGF-Flt Receptors B. Other Angiogenic Factors 11. Structure and Functions of VEGF/VPF A. Background B. Basic Structure of VEGF C. Functions of VEGF D. Expression of VEGF Gene E. Regulation of VEGF Gene Expression F. Subtypes of VEGF and Their Biological Significance G. Growth Factor or Growth Factor-like Sequences Closely Related to VEGF 111. Flt Rkceptor Gene Family: The Major Receptors for VEGF A. Background B. Structural Characteristics of the Flt Recepor Family C. Genomic Structure of the Flt Family D. Biological Functions of the Flt Family E. Signal Transduction of the Flt Family F. Expression of the F2t Family Genes IV. Regulation of Tumor Growth by the Suppression of VEGF-Flt Receptor System A. Suppression of Tumor Angiogenesis B. Control of Ascites Formation and Metastasis References

1. Introduction

A. NEWSYSTEM FOR THE REGULATION OF ANGIOGENESIS: VEGF-FLT RECEPTORS T h e vascular network in the body is crucial for the development and maintenance of a variety of normal tissues in higher organisms. Vascular formation (neovascularization) includes vasculogenesis and angiogenesis: vasculogenesis is blood vessel formation from endothelial precursor cells in the mesenchyme during the early stages of the embryo, and angiogenesis is the branching out of new vessels from preexisting small 28 1 ADVANCES I N CANCER RESEARCH, VOL. 67

Copyright Q 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

282

MASABUMI SHIBUYA

blood vessels. Vascular formation is important not only for the establishment of stable circular systems during embryogenesis in vertebrates but also for the rapid and transient angiogenesis under physiological conditions of the adult such as the formation of corpus luteum. Further, angiogenesis is widely known to be associated with various pathological conditions: inflammatory lesions, wound healing, endocrine diseases particularly diabetes mellitus, and growing tumors in vim. Blood vessels are considered to play a crucial role in the progression of cancer in vivo. ( 1 ) Tumor angiogenesis: solid tumors larger than 3 mm diameter were reported to require supplies of nutrients and oxygen through newly formed vessels for their rapid growth (Folkman, 1990). (2) Permeability of blood vessels: abnormal increase in vascular permeability may result in ascites or pleural effusion when metastatic tumor cells are growing in the abdominal or thoracic cavities. (3) Adhesion of tumor cells to vascular endothelial cells: cell-cell interaction of tumor cells that have invaded the vessels with the local endothelial cells appears to be one of the initial key steps for cancer metastasis through the bloodstream. These observations indicate that one of the possible therapeutic methods to suppress the progression of solid tumors in vivo is the blockage of blood vessel formation in tumor tissues and the suppression of abnormal permeability. To further understand the mechanisms underlying angiogenesis during pathological conditions, an extensive analysis of the physiological processes of vasculogenesis and angiogenesis at molecular levels obviously is essential (Risau et al., 1988; Wang et al., 1992). How similar are the molecular mechanisms that regulate vasculogenesis and angiogenesis? How are the angiogenic processes closely related o r different at normal and pathological conditions? To answer at least a part of these questions, studies have revealed a possible key system for angiogenesis, VEGF (vascular endothelial growth factor)/VPF (vascular permeability factor) and its receptor Flt family (Fig. 1). The purpose of this review is to discuss the following: (1) T h e VEGF-Flt receptor family system is utilized under normal angiogenic conditions basically in a paracrine manner. (2) VEGF-Flt appears to be one of the major signal transduction systems in reciprocal communication between endothelial cells and the surrounding parenchymal cells. (3) VEGF-Flt clearly is involved in a wide variety of tumor angiogeneses in vivo.

B. OTHER ANGIOCENIC FACTORS Several growth factors have been reported to carry angiogenic activity in vitro and/or in vivo: the fibroblast growth factor (FGF) family such as

ROLE OF VEGF-FLT IN ANGIOGENESIS

m-

283

Flt-1

KDRRUr-1

rrl

Flt-4

Endothelid cell growth Vascular permeability Macrophage migration N o d tissue vascularization Tumor vasculariaation

Glucose hansport Calcium efflux (Xenopus oocyte) FIG. 1. VEGF-Flt receptor system.

acidic FGF (FGF-l), basic FGF (FGF-2), Int-2, and Hst-1, as well as epidermal growth factor (EGF) and transforming growth factor (TGF)-or. All of these ligands mediate signals through receptor-type tyrosine kinases. However, these ligands have a broad spectrum of target cells and, therefore, are not specific to vascular endothelial cells. Neither acidic FGF nor basic FGF bears a typical signal peptide, thus, it is still an open question whether these major FGFs could be efficiently secreted out from the intact cells. Other members of the FGF family, Int-2 and Hst-1, contain signal peptides, but their gene expression in most normal tissues and cells is extremely low. Further, members of the EGF family such as EGF and TGFa independently do not show strong growth-stimulatory activity on most of the endothelial cells. In contrast with these growth factors, target cells for VEGF are basically endothelial cells. At histological levels the vascular system is well established as a multicellular organ consisting of endothelial cells, smooth muscle cells, pericytes, and some fibroblasts. Further, activated forms of macrophages and smooth muscle cells are also involved in the formation of inflammatory or degenerative vascular lesions. Many cytokines are also secreted from these cells, enabling their communication, but this chapter concentrates primarily on VEGF and its receptors.

II. Structure and Functions of VEGF/VPF A. BACKGROUND VEGF/VPF was originally isolated by Senger et al. (1983) as a factor regulating vascular permeability. In an attempt to identify a factor(s)

284

MASABUMI S H I B U Y A

that might be secreted from ascites-generating tumor cells and that might also cause an increase in the vascular permeability within the abdominal cavity, they purified from ascites and culture medium of guinea pig tumor cells a heparin-binding factor 40-45 kDa in molecular weight, which actually consisted of a dimer of two 20-23-kDa subunits. This factor was designated as vascular permeability factor since it increased the permeability of blood vessels when injected intracutaneously (Senger et al., 1983). Independently, another group (Ferrara and Henzel, 1989) found a new growth factor specific to endothelial cells in the conditioned medium of bovine pituitary folliculostellate cells by using an assay system for the growth of adrenally derived endothelial cells. This factor, named VEGF, did not stimulate the proliferation of fibroblasts nor epithelial cells, but caused an increase in the cell numbers of all types of endothelial cells examined so far. Surprisingly, however, independent molecular cloning and characterization of VPF and VEGF demonstrated that these two factors are essentially the same protein derived from a single gene (Leung et al., 1989; Keck et al., 1989; Ferrara et al., 1991b, 1992). When the original VEGF cDNA was compared with the VPF cDNA, VPF carried a 24-amino-acid insertion. As described later, this insertion was due to an alternative splicing, and a purified VPF was confirmed to have VEGF activity on endothelial cells (Connolly et al., 1989a). Likewise, VPF activity was detected in the VEGF molecule. Thus, to simplify the nomenclature in this article, VEGFNPF will be abbreviated hereafter as VEGF. €3. BASICSTRUCTURE OF VEGF

The VEGF gene has been isolated from a variety of animals: human, mouse, rat, bovine, and guinea pig (Leung et al., 1989; Keck et al., 1989; Senger P t al., 1990; Conn et al., 1990; Breier et al., 1992). Although the avian form of VEGF has not yet been molecularly cloned, isolation of two genes of the Flt receptor family from avian cDNAs strongly suggests that the VEGF-Flt receptor system also exists in birds (Eichmann et al., 1993). Genomic and cDNA analysis of the human VEGF gene revealed at least four subtypes of VEGF, i.e., VEGF,,,, VEGF16,, VEGFIR9,and VEGF,,, (the numbers indicate the amino acid residues in the mature form of VEGF protein) (Houck el al., 1991; Ferrara et al., 1991a) (Fig. 2). All subtypes of VEGF share the amino-terminal 14 1 amino acids, includ-

--

285

ROLE OF VEGF-FLT IN ANGIOGENESIS

exon 1-5

VEGF121

exon 8

exon7

VEGF165

exon6

VEGF189

I

115aa

24aa

17aa

exon 1-5

exon 7

exon6

PIGF-2

8

7

exon6 6

VEGF206

a

approx. 120aa

21aa

7 44aa

8 6aa

(149aa)

7

(17oaa)

8aa

FIG.2. Structure of the VEGF-PICF family. Subtypes of human VEGF and PlGF are indicated. Amino acid numbers for exons 1-5 in VEGF are outside the signal peptide. The amino acid numbers of PlGF in parentheses show the lengths of the entire coding sequences, including the signal peptide.

ing a signal peptide encoded by exons 1-5, and the carboxyl-terminal6 amino acids encoded by exon 8. T h e peptide derived from exon 7 has a basic feature and weak heparin-binding activity. Exon 6 present in both VEGFI89 and VEGF,,, encodes a highly basic stretch of 24 amino acids. VEGFZo6carries an additional short stretch designated as exon 6’; this stretch is derived from the contiguous downstream sequence of exon 6 due to the shift o! the exon 6 splicing donor site further downstream into the intron. Thus, it is not derived from an independent exon. There have been only a few reports about VEGFZo6so far, and the results from researchers including our group indicate that the major subtypes of VEGF in both normal and tumor tissues are of the 121-, 165-, and 189-amino acid-type molecules. As an exceptional case, placenta expresses VEGFI2, as the major type (Houck et al., 1991). While VEGF,,, has mild binding activity, VEGF,,, has strong binding activity to heparin or heparin-like molecules present on the cell surface, suggesting some different biological functions among the VEGF subtypes in vivo.

286

MASABUMI SHIBUYA

In the amino-terminal core region of VEGF encoded from exons 1-5, eight cysteine residues exist that contribute to dimer formation. Further, the amino-terminal 32-kDa molecule as a dimer generated from VEGF,,, by plasmin cleavage exhibited VPF activity, as well as VEGF activity (Houck et al., 1992). On the basis of these results, the fundamental biological activity needed for stimulation of endothelial cells is considered to be localized in this core region. Interestingly, this VEGF core is homologous to the amino-terminal two-thirds of PDGF-A and -B in terms of the amino acid sequences, including the distribution of eight cysteine residues and the exon-intron organization, clearly indicating that the VEGF and PDGFs belong to a supergene family (Keck et al., 1989). As in the case of PDGF and other growth factors, the VEGFs are also glycoproteins. This glycosylation was shown to be required not for the biological activity but for their stability and efficient secretion (Ye0 et al., 1991; Peretz et al., 1992). C. FUNCTIONS OF VEGF 1 . Stimulation of Proliferation and Tubular Formation of Endothelial Cells

VEGF stimulates cell growth in a variety of endothelial cells, such as human umbilical vein, bovine aorta, adrenal tissue, mouse brain, liver sinusoids of rats, etc. (Ferrara and Henzel, 1989; Connolly et al., 1989a; Gospodarowicz et al., 1989; Plouet et al., 1989; Yamane et al., 1994). Surprisingly, however, until now, none of the endothelial cell line established so far have been shown to be solely dependent on exogenously added VEGF for their growth, and hence all of the endothelial cells shown to proliferate in response to VEGF are primary culture cells. Sinusoidal endothelial cells of rat liver in particular show strict VEGF dependency when compared to other endothelial cells: they can neither grow nor survive in the absence of VEGF, irrespective of the presence of FGFs or high concentrations of serum, indicating that VEGF may act not only as a growth factor but also as a survival factor on certain types of endothelial cells (Fig. 3) (Yamane et al., 1994). Why do the endothelial cell lines lose their responsiveness to VEGF? One of the reasons for this could be due to a decrease in the expression of the VEGF receptors and the acquisition of the VEGF-independent growth-stimulatory mechanism, since the levels of fit-I and KDRIflk-I transcripts are very low in the endothelial cell lines examined so far (S. Yamaguchi and M. Shibuya, unpublished). VEGF stimulates tubular formation on endothelial cells in Matrigel

ROLE OF VEGF-FLT IN ANGIOGENESIS

287

b Cell No. x 10-2 ( a v d 4 q.) 4

Relative Cell No.

4 dayculture

zolll

' O 0i , 0

3

2

10

1

1

2

3

4

days in culture

VEGF

+

d- + -n+

Endo- Macrc- Epithet the1 phage

FIG. 3. Growth-stimulatoryactivity of VEGF on sinusoidal endothelial cells of rat liver (Yamane et al., 1994). (a) Morphological changes in the presence or absence of VEGF. (b) Time course of sinusoidal endothelialcell growth with or without VEGF (left). No response was observed on macrophage-like Kupffer cells or epithelial cells to VEGF (right).

and Collagen gel assays (Connolly et al., 198913; Goto et al., 1993). As expected, VEGF induces angiogenesis in vivo in both mammals and birds (Connolly et al., 1989a; Leung et al., 1989; Plouet et al., 1989; Wilting et al., 1992, 1993). Although the process of angiogenesis in vivo is known to be a complex one, accumulating evidence strongly suggests that VEGF is a direct angiogenic factor in vivo as well as in vitro. In addition, endothelial cells isolated from lymphatic vessels have been

288

MASABUMI SHIBUYA

shown to respond to VEGF for in vitro angiogenesis and induction of protease genes (Pepper et al., 1994). 2 . Increase in Vascular Permeability As Senger et al. (1983) demonstrated originally, VEGF bears permeability stimulatory activity on blood vessels as detected by Miles assay. This activity is equal to or even higher than that of other vascular permeability factors: the effective concentration of VEGF on vascular permeability is about 1 nM, which is 100-1,000-fold stronger than that of histamine and bradykinin (Connolly et al., 1989b; Connolly, 1991). Interestingly, histamine, a typical permeability factor, does not show any VEGF activity, whereas basic FGF, a growth factor on some types of endothelial cells, does not have any permeability activity (Connolly, 1991). In contrast, VEGF has both activities and thus is quite unique among growth and permeability factors. Permeability of VEGF can be blocked by the coinjection of anti-VEGF antibody, indicating that the permeability activity within the VEGF preparation is an innate property uf VEGF (Sioussat et al., 1993). Stimulation of permeability is thought to be mediated by signal transduction from the activated VEGF receptors (most likely Flt-1 and KDR/Flk- 1) toward cytoskeletal systems to enhance transcytosis and/or relaxation of cell to cell connection between endothelial cells (Collins et al., 1993). But the precise molecular mechanisms for this phenomenon are not yet clearly understood. 3 . Other Biolopcal Activities of VEGF on Endothelial Cells

One of the first steps in angiogenesis is the digestion of a portion of the intercellular matrix by proteases. VEGF has been reported to stimulate both the mRNA and protein levels of the interstitial collagenase gene in HUVEC (human umbilical vein endothelial cells) (Unemori g t al., 1992). On the other hand, plasminogen activator (PA) and PA inhibitor are up-regulated in response to VEGF in certain types of endothelial cells (Pepper et al., 1994), but not in HUVEC (Bikfalvi et al., 1991). The biological significance underlying the difference among endothelial cells is not yet understood. In the presence of VEGF, hexose transporter GLUT-1 gene expression is enhanced (Pekala et a/., 1990), and von Willebrand factor is released from Weibel-Palade bodies in endotheliai cells (Brock et al., 1991); the latter might play an important role in the initial step for the

ROLE OF VEGF-FLT IN ANGIOGENESIS

289

activation of the blood coagulatinn svstem in some pathological conditions, such as hypervascularity in diabetes mellitus- or abnormal vascularity in solid tumors. Ku et al., (1993) reported that, in response to VEGF, the coronary artery endothelial cells secrete nitric oxide (NO) (EDRF, endotheliumderived relaxation factor), resulting in dilatation of the coronary artery in the heart. Since vascular smooth muscle cells and cardiac myocytes were shown to express VEGF mRNA in hypoxic conditions (Ladoux and Frelin, 1993), these communication systems between endothelial cells and muscle cells in coronary artery using VEGF and NO appear to be crucial for the maintenance of appropriate amounts of oxygen and nutrition supplies in the heart. 4. Effects

of VEGF on Cells Other Than Endothelial Cells

Although the cell growth- and permeability-stimulatory activities of VEGF are restricted to endothelial cells, as a rare case, VEGF was reported to stimulate the migration of human peripheral blood monocytes (Clauss et al., 1990). In addition, an increase in procoagulant tissue factor activity and in the influx of Ca2+ ion after treatment with VEGF was also reported on monocytes (Clauss et al., 1990). Stimulation of the efflux of Ca2+ ion was observed in the human Flt-1 receptor overexpressing Xenopus oocytes in response to VEGF (De Vries et al., 1992), suggesting the existence of a signaling pathway toward Ca2+ ion channels in cell membrane from the VEGF receptors. Further, VEGF has been shown to induce differentiation in cultured osteoblasts (Midy and Plouet, 1994), raising the possibility that VEGF might also be involved in bone remodeling.

D. EXPRESSION OF VEGF GENE 1 , Embryogenesis

In early stages of embryogenesis such as E8 (8-day embryo) and E9 of rats, VEGF mRNA was shown to be relatively high in the trophoblast giant cells of placenta, but very low in the embryo. However, around E 11 the levels of VEGF mRNA increased in the thoracic region of the embryo, including the early stage of the heart, and peaked during E 11-El4 in the developing heart (Jakeman et al., 1993; Shifren et al., 1994).Then, VEGF mRNA levels decreased slightly in the heart during E14-EI6, but increased in the lung, brain, and kidney cortex. This pattern, i.e., higher

290

MASABUMI SHIBUYA

levels of VEGF mRNA in lung, kidney, heart, and brain, continued even after birth (Jakeman et al., 1993). At E15, the kidney epithelium surrounding the glomerulus was also reported to strongly express VEGF mRNA (Breier et al., 1992). 2. Normal Adult Tissues Like embryonal stages, various adult tissues such as lung, kidney, adrenal, heart, and ovary express easily detectable levels of VEGF mRNA (Ferrara et al., 1991c; Berse et al., 1992). Cytologically, alveolar epithelial cells in lung, podocytes, and mesangial cells in glomeruli (Brown et al., 1992a, 1993a; IiLjima et al., 1993), culured glomerular endothelial cells (Uchida et al., 1994), adrenal cortex, myocytes in heart, and pas distalis cells in hypophysis were shown to contain relatively high levels of VEGF rnRNA. Since the endothelial cells in these adult tissues are considered to be at the stage of growth arrest, a very important question concerns the biological role of VEGF in these normal tissues. An attractive and possible function of VEGF in normal adult tissues such as lung and kidney may be the maintenance of the proper permeability and the integrity of vascular endothelial cells. In the adult female, rapid neovascularization is known to occur during corpus luteurn formation, endometrial growth of the uterus, and development of the placenta (Sharkey et al., 1993; Dissen et al., 1994). Interestingly, preceding these processes, a strong and transient expression of the VEGF gene at both mRNA and protein levels was observed in comulus oophorus cells and luteal cells in ovary and endometrial cells in the uterus (Phillips et al., 1990; Ravindranath et al., 1992; Shweiki et al., 1993; Charnock-Jones et al., 1993), clearly indicating the involvement of VEGF in this neovascularization. Experiments have shown that the induction of VEGF rnRNA with estrogen in rat uterus is well correlated with the increase in capillary permeability and angiogenesis in this tissue (Cullinan-Bove and Koos, 1993), and granulosa cells in culture were shown to express VEGF mRNA in a hormone-dependent manner (Garrid0 et al., 1993; Yan et al., 1993). In addition, activated macrophages (oil-induced peritoneal macrophages), cultured smooth muscle cells, and retinal pigment epithelial cells were also reported to express VEGF mRNA (Ferrara et al., 1991b; Berse et al., 1992; Adamis et al., 1993). Therefore, a variety of mechanisms may exist in adult animal tissues for the maintenance of the steady state level or the modulation of VEGF gene expression.

ROLE OF VEGF-FLT I N ANGIOGENESIS

29 1

3. Expression of VEGF mRNA in Abnormal Cells in Pathological Conditions and in Tumor Cells Angiogenesis is well known to take place in the later stages of wound healing. Keratinocytes near the wounded lesion (activated condition?) have been shown to transiently express VEGF mRNA for the induction of endothelial cell growth (Brown et al., 1992b). I n ovulation-induction therapy, female patients occasionally have ovary hyperstimulation syndrome, which causes ascites with proteins released from vessels and, rarely, pleural and cardiac effusion. McClure et al., (1994) have shown that these ascites contain high levels of VEGF, strongly suggesting that the major cause of this syndrome is abnormal production of VEGF by hormonal therapy. Abnormal angiogenesis in retinas of patients with diabetes mellitus and in rheumatoid synovial tissues, as well as hypervascularity in the enlarged thyroid tissues of patients with Graves’disease, has been shown to be correlated with the abnormal induction of VEGF gene expression (Aiello et al., 1994; Miller et al., 1994; Fava et al., 1994; Koch et al., 1994; Yamazaki et al., 1994). Thus, in many cases of pathological angiogenesis, the VEGF-Flt receptor system appears to be deeply involved (Fig. 4). A number of tumor cells in humans and other mammals have been demonstrated to express higher levels of VEGF mRNA when compared with that present in the surrounding normal tissues. These tumors include glioblastoma multiform (GBM) (Plate et al., 1992; Shweiki et al., 1992; Berkman et al., 1993), meningioma, gastric cancer (Brown et al., 1993b), renal carcinoma (Brown et al., 1993a; Takahashi et al., 1994),

Epithelial cells, Mesenchymal cells, p53-mutation

EGF

.

EGF

TSH low 02

Estrogen Ca-ionophore TPA

0 - m 0

Imrecrrein permeability Secretioncd collagemrsr Tubular tormation

etc.

FIG. 4. Induction of VEGF under various conditions.

292

MASABUMI SHIBUYA

colon cancer (Berse et al., 1992), ovarian cancer (Olson et al., 1994), etc. (Dvorak et al., 1991). GBM is characterized by high tumor angiogenesis, focal necrosis, local brain edema, and rapid growth of tumor cells. Most of the primary GBM express high amounts of VEGF transcripts, and the levels of expression have good correlation with the levels of malignancy (Berkman et al., 1993). In terms of genetic alteration, about 40% of GBM carries gene amplification of wild-type o r structurally mutated EGFR gene (Libermann et al., 1985; Yamazaki at al., 1988, 1990; Humphrey et al., 1990), and more than 30% of GBM aquires a p53 mutation in the process of carcinogenesis. Therefore, reports showing an increase in VEGF mRNA levels in GBM after stimulation with EGF and hyperinduction of the VEGF gene through the C-kinase pathway in mutant p53-carrying tumor cells seem to be particularly interesting because of the possible connection between alteration of cellular oncogenes or tumor suppressor genes in tumor cells and hyperinduction of the VEGF mRNA. In many cases of GBM, the expression of basic FGF was also reported to be enhanced (Takahashi et al., 1990), suggesting that basic FGF is involved in tumor angiogenesis of GBM. However, in rat gliomas, which is a model system for human GBM, only VEGF but not the basic FGF mRNA was found to be increased (Plate et al., 1993). Further, basic FGF has been reported not to carry vascular permeability activity. Taken together, VEGF may be the major angiogenic and vascular permeability factor in GBM and in most of the other solid tumors (Fig. 4). In addition to tumor tissues, a variety of tumor-derived cell lines, for example, A431, U937, mouse sarcoma 180, and AtT20, express VEGF mRNA and produce VEGF proteins at variable levels (Plouet et al., 1989; Rosenthal et al., 1990). One important characteristic of VEGF production from tumor cells is that, unlike other growth factors such as FGF and TGFa secreted from tumor cells, VEGF appears to be unable to create an autocrine loop for cell growth; introduction of VEGF expression vector to culture cell lines does not stimulate the growth rate in uitro, but stimulates tumor growth in nude mice due to the enhancement of angiogenesis in u k o (Ferrara et al., 1993). In addition, most of the human tumors examined so far did not express thept-l gene, a VEGF receptor, at detectable levels by Northern analysis (Shibuya et al., 1990). Thus, these data are consistent with the idea that a vast majority of tumor cells produce VEGF for the induction of tumor angiogenesis, but not for growth stimulation. T h e question that remains to be answered is whether any angiogenic factors other than VEGF contribute to angiogenesis in tumors, and if so, how crucial is their role in tumor angiogenesis?

ROLE OF VEGF-FLT IN ANGIOGENESIS

293

4 . Expression of VEGF in Tumor Cells Derived from Endothelial Cells

Since VEGF is a growth factor specific to endothelial cells, it seems possible that VEGF and its receptors can be activated as oncogenes in these cells. VEGF mRNA was reported to be relatively high in human brain hemangioblastomas (Morii et al., 1993), indicating a rare case of autocrine stimulation, although this has not yet been confirmed at protein levels. On the other hand, chemical carcinogen-induced murine angiosarcoma showed very low levels of VEGF and its receptor mRNAs (S. Yamaguchi and M. Shibuya, unpublished); hence, in the latter case the tumor cells might have acquired a growth-promoting mechanism independent from the VEGF-Flt receptor family system. Kaposi sarcomas in AIDS patients are known to be hypervascularized, edematous lesions. Since Kaposi sarcoma cells represent some features of endothelial origin, it is important to ask whether the abnormally proliferating vascular endothelial cells within these tumors are Kaposi sarcoma cells or normal cells of host origin. High levels of VEGF mRNA in Kaposi sarcoma shown by Weindel et al., (1992) suggest a paracrine stimulation of normal endothelial cell growth. E.

REGULATION OF

VEGF GENEEXPRESSION

Analysis of the human VEGF genomic DNA revealed that the promoter region does not carry a typical TATA box, but has six GC boxes for SP1-binding sites and some AP-1- and AP-2-binding sites. On the basis of these data, VEGF gene expression is suggested to be enhanced with serum, phorbol ester TPA, and cyclic AMP. As expected, the VEGF gene was up-regulated in vascular smooth muscle cells after treatment with TPA (Tischer et al., 1991). Further, listed here are other chemicals, growth factors, and certain culture conditions that have been reported to induce VEGF mRNA: calcium ionophore A2318’7, cyclic AMP, and a condition for induction of cell differentiation (Claffey et al., 1992); hypoxia on a glioblastoma cell line and heart ventricular myocytes (Schweiki et al., 1992; Ladoux and Frelin, 1993); EGF on glioblastomas (Goldman et al., 1993); stimulation of the protein kinase C pathway in the presence of p53 tumor suppressor gene mutation (Kieser et al., 1994); and hormones such as estradiol on rat uterus tissue and cultured ovarian granulosa cells (Fig. 4) (Garrido et al., 1993; Yan et al., 1993). TGFP and prostaglandin also induced VEGF mRNA in epithelial cells and osteoblasts, respectively (Pertovaara et al., 1994; Harada et al., 1994). Among these, one of the

294

MASABUMI SHIBUYA

most interesting phenomena may be VEGF induction at hypoxia. Although this response was originally found in a tumor cell line, the fact that normal cells also bear this activity strongly suggests that up-regulation of VEGF under low 0, conditions is not an abnormal response in a transformed situation, but a fundamental biological response in a variety of normal cells in higher organisms. Erythropoietin (Epo) gene is known to be up-regulated in low 0, conditions. T h e pattern of the induction of VEGF gene expression in a glioblastoma cell line and that in a hepatoma-derived cell line Hep3B have been shown to be very similar to that of Epo gene expression in terms of the time course, suppression by CO (carbon monoxide), induction by cobalt ion, and stabilization of mRNA by cycloheximide (Goldberg and Schneider, 1994). Therefore, molecular mechanisms for the detection of low 0, conditions and for the signal transduction to modulate the expression of several critical genes such as Epo and VEGF may be highly conserved within normal cells and tissues. F. SUBTYPES OF VEGF AND THEIR BIOLOGICAL SIGNIFICANCE Among the major three subtypes of human VEGF, VEGF,,, does not contain a basic domain, whereas the other two subtypes, VEGF,,, and VEGF,,,, carry the heparin binding basic domain. PDGF, a growth factor distantly related to VEGF, has been shown to carry two subtypes (PDGF short form and long form): one does not contain a basic domain, and the other bears a basic domain that corresponds to exon 6 in the VECF gene. Thus, structurally, and also probably functionally, VEGF,,, and VEGF,,5,189 may correspond to the short form of PDGF and the long form of PDGF, respectively. Since VEGFI6,, VEGF,,,, and PDGF long form have the property to bind to the cell surface, the existence of both soluble and cell surface-bound forms in the PDGF-VEGF supergene family strongly suggests that two forms (free and cell-bound forms) of VEGF are necessary for the accomplishment of their roles in physiological angiogenesis (Park et al., 1993). What then is the biological role of the cell-bound form of VEGF? At least three possibilities can be considered: (1) the cell-bound form of VEGF has some additional biological function(s) when compared to that of the soluble VEGF; (2) the cell-bound form of VEGF is essentially a stored VEGF and is used after release from the cell surface (or intercellular matrix) by partial cleavage with proteases (Houck et al., 1992); and (3) both the soluble and cell-bound forms of VEGF retain almost the

ROLE OF VEGF-FLT I N ANGIOGENESIS

295

same biological functions, and the latter form is used in a cell to cell interaction between endothelial cells and the surrounding epithelial or mesenchymal cells to induce angiogenesis in the appropriate direction and at an appropriate amount. A similar situation was reported in the case of SCF (stem cell factor) and its receptor tyrosine kinase c-Kit, in which the membrane-bound form of SCF bearing a transmembrane domain due to an alternative splicing appears to be used as a guide for the proper migration of a certain type of c-Kit-positive cells in embryogenesis. G. GROWTHFACTOR OR GROWTH FACTOR-LIKE SEQUENCES CLOSELY RELATED TO VEGF 1 . Placenta Growth Factor (PlGF) A cDNA encoding a VEGF-related protein designated PlGF has been molecularly cloned from the human placenta cDNA library (Maglione et al., 1991). PlGF-1 consists of 149 amino acids, including a signal peptide. I n the core region encoded from exons 1-5 in VEGF, the structure between VEGF and PlGF is about 40% identical at the amino acid level, and all eight cysteine residues in this region are conserved in both VEGF and PlGF. A new subtype of PlGF, PlGF-2, which carries a highly basic region of 2 1 amino acids corresponding to exon 6 in VEGF, was isolated (Fig. 2): PlGF-2 is expressed 2-3-fold more than PlGF-1 in normal placenta. This PlGF basic region was confirmed to be derived from a single exon in the PlGF genome; thus, PlGF-1 and PlGF-2 are produced by an alternative splicing mechanism (Maglione et al., 1993b; Hauser and Weich, 1993). PlGF was also shown to have growth-stimulating activity on endothelial cells like VEGF, but a precise comparison between the activities of PlGF and VEGF has not yet been carried out. Although the VEGF gene is expressed in a variety of normal tissues, PlGF transcripts are detectable by Northern analysis only in placenta among the normal adult tissues examined. Further, embryonal tissues expressed very low levels of PlGF. However, a few tumor cells such as hepatoma-derived HepG2, choriocarcinoma, and renal cell carcinoma do express the PlGF gene (Maglione et al., 1993b; Takahashi et al., 1994). These observations suggest a role for PlGF in tumor angiogenesis in certain tumors, but analysis at protein levels remains to be done. PlGF mRNA contains a short open reading frame at the 5'-noncoding sequence, and the removal of this region enhances the efficiency of protein production in a baculovirus system (Maglione et al., 1993a). T h e

296

MASABUMI SHlBUYA

biological meaning of this short ORF in PlGF mRNA is not yet clear. PlGF has been reported to bind Flt-1 but not KDR/Flk-1 (Kendall et al., 1994). 2 . VEGF-Related OV (O$ Virus) Genes Encoded in Pox Viruses Among the strains of Pox virus, Orf viruses, which infect sheep, goat, and, rarely, humans, NZ-2 and NZ-7 strains are known to induce local angiogenesis and edema at the infected lesions. Studies have revealed that the open reading frames encoding VEGF-related proteins exist in the virus genome DNA (Lyttle et al., 1994). Orf-VEGFs present in NZ-2 and NZ-7 strains are closely related and 41% identical to each other at the amino acid level. T h e similarity between Orf-VEGFs and human VEGF or PlGF is about 16-2796 at the amino acid level. Orf-VEGF genes are considered to be derived from a cellular gene related to VEGF in vertebrates; therefore, the putative Orf-VEGF may share a unique gene family with VEGF and PlGF regulating endothelial cell growth and differentiation in animals. 111. Flt Receptor Gene Family: The Major Receptors for VEGF

A. BACKGROUND It is now widely accepted that a peptide growth factor like VEGF binds a specific receptor or receptor complex expressed on the cell surface and activates them to generate mitotic signals. After the isolation of VEGF, radioisotope-labeled VEGF was used as a tool to search the specific receptorts) on endothelial cells. From these examinations, several types of endothelial cells have been shown to express high-affinity binding sites of approximately 180 kDa on their cell surface: after crosslinking, a complex of approximately 230 kDa including the 45-kDa VEGF molecule was detected (Vaisman et al., 1990; Plouet and Moukadiri, 1990; Bikfalvi et al., 1991; Myoken et al., 1991; Olander et al., 1991; Gitay-Goren et al., 1992). However, the real structure and the enzymatic activities of the “receptors” were not elucidated. In an attempt to isolate new receptor-type tyrosine kinases that might be involved in carcinogenesis as activated oncogenes like EGF receptor and c-ErbBZ/Neu, we screened a human genomic DNA library with a tyrosine kinase sequence as a probe at very low stringency hybridization conditions (Matsushime et al., 1987). We obtained a DNA fragment containing an exon that encodes a portion of a new receptor-type tyrosine

ROLE OF VEGF-FLT I N ANGIOGENESIS

297

kinase. Unexpectedly, however, this gene was not overexpressed in any human tumor cell lines derived from various tissues, whereas it was found to be expressed, although at low levels, in most normal tissues (Shibuya et al., 1990). An 8-kb cDNA of this gene isolated from a normal human placenta cDNA library encoded a novel receptor-type tyrosine kinase distantly related to the structures of c-Fms (colony stimulating factor- 1 receptor), c-Kit (stem cell factor receptor), and PDGF receptor. Thus, this gene was designated as fit @t-1) Vms-like tyrosine kinase) (Shibuya et al., 1990). The fit-1 gene was molecularly cloned from rodents (Finnerty et al., 1993; Yamane et al., 1994). After the isolation of thefit-1 gene, a receptor gene closely related to fit-1 (named KDR) was isolated from a human endothelial cell cDNA library (Terman et al., 1991, 1992) and from a mouse fetal liver cDNA library @k-1 gene) (Matthews et al., 1991; Sarzani et al., 1992).Since KDR andfik-1 are derived from the same gene in different species of vertebrates, this gene is shown as KDRlfik-1 in the following section. Further, another receptor gene related to the flt receptor gene, designated fit-4, was isolated from a human and murine cDNA library (Pajusola et al., 1992; Galland et al., 1993; Finnerty et al., 1993).To simplify the nomenclature, I will use the term “thefit gene family,”which includesfit-related receptor genes. Flt- 1 and KDRIFlk- 1 receptors have been demonstrated to bind VEGF at high affinities (De Vries et al., 1992; Quinn et al., 1993; Millauer et al., 1993; Seetharam et al., 1994). At this moment thefit gene family consists of three members, thefit-1, KDRIfik-1, andfit-4 genes (Fig. 1). A clustering of the fmslfit supergene family in the three loci of mammalian chromosomes suggests that no other members probably exist in the flt gene family (Figs. 5 , and 6) (Satoh et al., 1987; Matthews et al., 1991; Aprelikova et al., 1992; Rosnet et al., 1993). B. STRUCTURAL CHARACTERISTICS OF THE FLT RECEPOK FAMILY

T h e human Flt-1 receptor tyrosine kinase consists of 1338 amino acid residues divided into three portions, an extracellular domain, a transmembrane domain, and a cytoplasmic domain carrying a tyrosine kinase domain. I n two-thirds of the extracellular domain, the localization of most cysteine residues completely matches that of the FmsIKitlPDGFR. These cysteines appear within a gap of every 50-60 amino acids and have been suggested to form five intramolecular disulfide bonds, al-

298

MASABUMI SHIBUYA

4qll-ql3

5q33-q34

13ql2-ql3

IF3 flt- 7

f k -2 / f k 3

FIG.5 . Chromosomal tocation of thefintlflt supergene family in humans.

though the fourth loop may not be complete due to the lack of one cysteine around residue 360 in Flt and Fms/Kit/PDGFR. I n addition, the Flt-1 receptor contains another stretch of 220 amino acids between residue 550 and the transmembrane domain. Since this Flt-specific region also carries four cysteines within a gap of every 50 amino acids, and since each possible loop formed by a disulfide bond bears a strong resemblance to the loops in the immunoglobulin (Ig) molecules, Flt-1 could be referred to as a 7-Ig-type receptor, whereas Fms/Kit/PDGFR could be

-1

+

cis duplication

mq trans duplication

FIG. 6. Hypothetical scheme of phylogenetic evolution of thefmslflt supergene family (Rosnet et nl., 1993). 5-Ig and 7-Ig indicate thefmr family and theflt family, respectively.

ROLE OF VEGF-FLT IN ANGIOGENESIS

299

referred to as 5-Ig-type receptors (Fig. 7). It is of interest to determine how the Flt-specific 220-amino acid region was generated during the evolution of genomic DNA in animals. A long kinase insert, 60-70 amino acids long, at the same position within the tyrosine kinase domain is a common feature among Flt-l/ KDR(F1k-1)/Flt-4 and Fms/Kit/PDGFR. Interestingly, however, the Flt family kinase inserts do not have the Tyr-X-X-Met motif, which is well conserved in the Fms family kinase insert and is a major autophosphorylation site important for signal transduction through the binding and activation of phosphatidylinositol 3-kinase (Fant et al., 1992; Valius and Kazlauskas, 1993). Furthermore, even in the entire cytoplasmic domain of the Flt family no Tyr-X-X-Met motif was present, strongly suggesting that the signal transduction pathway from the Flt receptors is different from that generated from Fms/Kit/PDGFR. Among the Flt family, Flt-1 and KDR/Flk-1 appear to be structurally more closely related on the basis of the similarities of the kinase insert

b

YTAV

YIIP

FIG.7. Structure of Flt-1 tyrosine kinase and its comparison with one of the Fms family, PDGFRD. All of the tyrosines (Y) indicated on PDGFR are autophosphorylation sites.

300

MASABUMI SHIBUYA

sequences (about 50% identity) (Table I) and the carboxyl-terminal short stretch Ser-Ser/Thr-Pro-Pro-[hydrophobic amino acid]. Supportingly, VEGF has been reported to bind both Flt-1 and KDRIFlk-1, but there has been no report showing high-affinity binding of VEGF to Flt-4. Within the Flt-1 cytoplasmic domain, Tyr residue 1169 (Tyr-Ile-ProIle) and Tyr 12 13 ('Tyr-Val-Asn-Ala) could be possible binding sites for the signal transducer SHPTP (SH2-containing phosphotyrosine phosphatase) and Grb2, respectively, although phosphorylation of these tyrosines on Fit-1 has not yet been examined. C. GENOMIC STRUCTURE OF THE Flt FAMILY

A partial genomic structure of theflt-4 gene was reported, and the organization of exons and introns in the kinase domain was found to be highly conserved amongjt-4, fm and kit genes (Pajusola et al., 1993). Further, we detected a few additional introns in the 5'-region of thefit-1 gene compared to that of thefm and kit genes, suggesting a phylogenetic order in the formation o f j t and fms family genes (K. Kondo and M. Shibuya, unpublished). In addition to the full size mRNAs, thefit-1 gene expresses high levels of two short mRNAs in placenta, 293E1, and a few other cell lines (Shibuya et al., 1990). Most of these shortpt-1 mRNAs are considered to

.4!dINO A C I D BETWEEK

FLI- 1

TABLE I HOMOLOGY 1.K T H E TYROSINE KINASEDOMAIN A N D &HER

RECEPTOK- O R NONRECE&TOR-'rYPE

TYROSINE KINASES Ty rosine

Amino half.

kinase

(57)

KDFIFlk-I FLt-4

78 80 57 59 5. 7 62 26 32 30 32

v-Fms

PDGFRa PDGFRP v-Kit

v-ErbB V-ROS v-Src v-Fps

Flt-I residues 813-929.

', Fit- 1 I-esidues930-994.

Fit-I residues 995-1 152.

Inserth (%)

Carboxyl half' (%)

51 11

78 81 53 58 53 58 42 43 38 41

3

13 13 11

ROLE OF VEGF-FLT I N ANGIOGENESIS

30 1

encode the amino-terminal half of the Flt-1 protein, which consists of Flt- 1 1-656 residues and an additional 3 1 amino acids derived from the intron sequence (Kendall and Thomas, 1993). The amount of thisJEt-1 short mRNA is 5-10-fold higher than that of the full length mRNA in placenta, thus, the putative small Flt-1 protein may have some biological function(s) in the development and maintenance of this tissue. TheJEt-4 gene has been shown to have two types of transcripts that bear different carboxyl-terminal sequences due to an alternative splicing (Pajusola et al., 1993). The significance of this difference has not yet been elucidated. OF THE FLT FAMILY D. BIOLOGICAL FUNCTIONS

T h e biological activity of VEGF is mostly specific to endothelial cells, and VEGF stimulates cell proliferation, permeability, tubular formation, and production of proteases and other small molecules in endothelial cells. Migration of macrophages and differentiation of osteoblasts have also been shown to be stimulated with VEGF (Clauss et al., 1990; Midy and Plouet, 1994). Since VEGF-binding molecules at high affinity so far are limited to Flt-1 and KDR/Flk-1 receptors, it seems most likely that the preceding biological activities are transmitted by the activation of these two VEGF receptors. We cannot rule out the possibility, however, that Flt-4 and another as yet unidentified receptor also contribute to VEGF-induced signal transduction. T h e sizes of the VEGF receptors on macrophages are similar to those on endothelial cells (Shen et al., 1993), whereas the binding molecules on fibroblasts are shorter (about 120- 150 kDa) than Flt-1 and KDR/Flk-1; thus, the characteristics of these cell surface molecules are not clear at this moment. Rat sinusoidal endothelid cells express high levels offlt-I and KDRIJEk-I mRNA, although the latter appears to be severalfold less than the former (Yamane et d., 1994). These cells have a strong dependency on VEGF for cell growth and maintenance, and the morphologies of these cells in the presence or absence of VEGF were different from each other (Yamane et al., 1994; T. Takahashi and M. Shibuya, unpublished). Therefore, the activated Flt- 1 and KDR/FIk- 1 receptors may modulate the cytoskeletal organization of the sinusoidal endothelial cells. In the Xenopus oocyte system, both exogenously introduced wild-type Flt-1 and a mutant Flt-1 carrying a 67-amino acid deletion at the carboxyl region were able to stimulate calcium mobilization (De Vries et d., 1992). Thus, the Flt-1 carboxyl sequence seems dispensable for this particular signaling pathway. It will be of interest, as well as important, to

302

MASABUMI SHIBUYA

examine part by part the cytoplasmic domain of the Flt family of molecules to determine which region bears which biological function of VEGF. E. SIGNAL TRANSDUCTION OF THE FLT FAMILY In general, signal transduction from receptor-type tyrosine kinases is proposed to be as follows (Schlessinger and Ullrich, 1992; Egan et al., 1993): (1) dimerization and oligomerization of the receptors through ligand binding; (2) activation of tyrosine kinases and autophosphorylation of the receptor; (3) association of signal transducers such as adaptors (Shc and Grb2), PI-3 kinase, PLCy, and other tyrosine kinases (Src, etc.) to the phosphotyrosine-containing regions through the SH2 domain; (4) tyrosine phosphorylation of these associated molecules; (5) translocation of the Grb2-Sos (GTP-GDP exchange factor to Ras) complex to the cell membrane; (6)activation of Ras to the GTP-bound form; and (7) activation of MAP kinase cascade. I . Oligomenultion of Flt Receptors

When the Flt-1 receptor is expressed on COS cells and NIH3T3 fibroblasts, dimer and oligomer formation of the receptor was observed in response to VEGF (De Vries et al., 1992; Seetharam et al., 1994). These results suggest that association between Flt- 1 and VEGF is sufficient for the structural activation of the receptor, although heparin has been reported to stimulate the binding of VEGF to its receptor(s) (GitayGoren et al., 1992; Tessler et al., 1994). Oligomerization of the KDR/Flk-l receptor expressed in COS cells in the presence of VEGF has not yet been clearly demonstrated (Millauer et al., 1993). 2 . Autophosphorylataon of the Flt Family

T h e level of autophosphorylation of the Flt-1 receptor on tyrosine residues after stimulation with VEGF was found to be very low (or barely detectable) in transient expression of Flt-1 in Xenopus oocytes or COS cells (De Vries et al., 1992), in stable expression of Flt-1 in NIH3T3 cells (Seetharam et al., 1994), and in rat sinusoidal endothelial cells in liver (Yamane et al., 1994). These results suggest that the autophosphorylation activity of Flt-1 itself is very low or the turnover of the tyrosine phosphorylation on Flt-1 is quite rapid within the cells. On the other hand, the autophosphorylation of KDR/Flk- 1 receptor expressed in COS cells was easily detectable in response to VEGF. Since the amino acid sequences in the tyrosine kinase domain are highly homologous (about 78% identical) between Flt- 1 and KDRIFlk-1, it should be very interesting to elucidate the mechanism and the biological signifi-

ROLE OF VEGF-FLT IN ANGIOGENESIS

303

cance of the difference in the autophosphorylation levels in these two VEGF receptors (Waltenberger et al., 1994). 3 . Tyrosine Phosphorylation of Cellular Substrates and Activation of M A P Kinase Cascade by the Flt Family

In the Flt- 1 overexpressing NIH3T3 cells, tyrosine phosphorylation of the GAP (GTPase activating protein of Ras) complex and PLCy was observed in response to VEGF. Unexpectedly, however, unlike many other receptor tyrosine kinases such as EGFR, insulin-R, NGFR, and CSF- l-R (c-Fms), tyrosine phosphorylation of the Shc adaptor protein or MAP kinase as a marker for the activation of the Ras-MAPK pathway was very low in Flt-1 NIH3T3 cells (Seetharam et al., 1994). Since a variety of receptor tyrosine kinases, particularly wild-type and autophosphorylation-site-minus mutant EGFRs, utilize Shc phophorylation for signal transduction toward Ras (Gotoh et al., 1994; Batzer et al., 1994), poor phosphorylation activity of Flt-1 on Shc seems unique among various receptors. Furthermore, induction of c-fos or c-myc was very low, and DNA synthesis was not stimulated in the Flt-1 overexpressing NIH3T3 cells in the presence of VEGF. On the other hand, in rat sinusoidal endothelial cells, which express fit-1 and KDRlftk-1 mRNA, MAP kinase was dramatically activated upon stimulation with VEGF, although the phosphorylation of Shc was still at low levels (Seetharam et al., 1994). Thus, if one assumes that DNA synthesis in endothelial cells is mediated at least in part by the activation of Flt-1, this signaling appears to be incomplete in the fibroblast background (Fig. 8). So far, the information regarding signal transduction from activated KDR/Flk-1 (and Flt-4) is very limited: only one report describes the stimulation of DNA synthesis in KDR/Flk-1 overexpressing NIH3T3 cells in response to VEGF (Millauer et al., 1994). However, it is not yet clear whether this induction of DNA synthesis is enough for regular cell cycles, and whether cell mitosis is complete in KDR/Flk-1 expressing cells. I n addition, the CSF-l/Flt-4 chimeric molecule could not induce DNA synthesis in endothelial cells in response to CSF-1 (Pajusola et al., 1994). Do endothelial cells have their specific signal transduction pathway from the Flt receptor family for cell growth? The fact that no cells or cell lines other than endothelial cells expressing Flt-1 or KDR/Flk-1 were reported to grow in response to VEGF strongly suggests the existence of an endothelial-specific signaling mechanism. However, it might also be possible that heterodimer or heterooligomer formation among the activated Flt- 1, KDR/Flk- 1, and Flt-4 is essential for signaling, as in the case

304

MASABUMI SHIBUYA

EGFR

n

t

Shc

MAPK

t MAPK

FLT- 1

1

P/Y 0 Shc

/PLCr

r'

1

t MAPK

FIG. 8. Model for signal transduction from EGFR and Flt-I.

of TGFP receptor signaling, which utilizes receptor heterodimer formation (Wrana et al., 1994). Further analysis on signal transduction in endothelial cells is necessary for understanding of the fundamental characteristics of the endothelial cells and blood vessel formation. F.

EXPRESSION OF THE Flt FAMILY GENES 1. E m bryogenesis

Most ofthe analyses on this stage were carried out by an in situ hybridization technique using human, mouse, and avian embryonal tissues (Yamaguchi et al., 1993; Millauer et al., 1993; Eichmann et al., 1993; Peters et al., 1993; Kaipainen et al., 1993; Quinn et al., 1993). Basically the fZt family gene transcripts are localized in vascular endothelial cells from the stage of early embryogenesis (blood island formation) to the middle stage, when the vascular systems are extensively developed in

ROLE OF VEGF-FLT IN ANGIOGENESIS

305

many tissues. In the late stage,jt family gene expression gradually decreases. These results are consistent with studies on the localization of VEGFbinding sites by using [125I]VEGF (Jakeman et al., 1993), in which the binding sites of radiolabeled VEGF (thus, most likely the summation of Flt- 1 and KDRIFlk-1) are detectable at very early stages of blood vessel formation (hemangioblasts) and restricted to the vascular system in the all-embryonal tissues. Northern blot analysis further confirmed the high expression of the j t family from mouse E9.5 to E18.5 in total RNAs obtained from the whole embryo (Matthews et al., 1991). These findings stongly suggest that the VEGF-Flt receptor system is involved in both vasculogenesis and angiogenesis during the early stages of embryogenesis. Although gene expression of the fit family is essentially localized in endothelial cells, some differential pattern in the expression of the j t members has been reported. In the case of birds, the Quekl gene, related to marnmalianjt-4, is expressed in the mesoderm from the onset of gastrulation, whereas the mRNA of the Quek2 gene, corresponding to mammalian KDRIjk-I, is first detected in early endothelial cells (Eichmann et al., 1993). I n human embryo, bothjt-1 and KDRI’k-1 were expressed in small vessels in the epicardium; however, in the coronary endothelium only j t - 1 but not KDRIjk-1 mRNA was clearly detected. Further, fit-1 and KDRIjk-1 transcripts were positive, butjt-4 was negative in capillaries of the myocardium. Similarly, some differences in expression among the j t family were observed in other tissues (Kaipainen et al., 1993). I n addition, the’t-4 mRNA is higher in the cortical plate, intermediate zone, and cerebellum in human fetal brain (Pajusola et al., 1992). Thus, the fit family might have some additional, although not proven, role in the development of the central nervous system. 2. Adult Tissues

I n the adult stage, thejt-1 mRNA is detectable in most of the normal tissues: particularly at higher levels in placenta, lung, kidney, and brain (Shibuya et al., 1990). These mRNAs are thought to be expressed in vascular endothelial cells since, for example, in the liver tissue only the sinusoidal endothelial cell fraction but not the hepatocyte fraction expresses very high levels o f j t - I and KDRI’k-1 mRNAs (Yamane et al., 1994). These results are consistent with the localization of I125IlVEGFbinding sites on vascular endothelial cells (Jakeman et al., 1992). On the other hand, in rat liver, only hepatocytes express VEGF mRNA, but the endothelial cell fraction does not show any VEGF gene expression. Very

306

MASABUMI SHIBUYA

interestingly, another ligand-tyrosine kinase receptor system, i.e., HGF (hepatocyte growth factor), and its receptor c-Met are expressed in liver in an opposite direction: HGF mRNA is detectable only in the sinusoidal endothelial cell fraction, whereas c-Met transcripts are restricted to the hepatocyte fraction (Yamane et al., 1994) (Fig. 9). Therefore, two reciprocal communication systems, i.e., between parenchymal cells and endothelial cells, interdependent on each other may exist in liver tissue utilizing the VEGF-Flt receptor system (Fig. 10). It seems very likely that many other tissues also contain similar paracrine loops of reciprocal communication systems, including the VEGF-Flt receptors (Fig. 11). During transient angiogenesis in the hormone-regulated reproductive organs of females, [125I]VECF-binding sites were reported to be localized only on the endothelial cells, again supporting the strict regula-

FIG. 9. Dif.ferenrial expression of the VEGF-Flt receptor family and HGF-Met recepgenes at rat liver (Yamane rt nl., 1994). About 10 pg of total RNA was obtained from whole li\,er or the fractionated cells and analyzed by Northern blotting. Abbreviations: XI'<;. rumparenchymal cell t-raction (9'2% are siiiusoidai endothelial cells): N - 1 and -2, parench~nialcell fractions (more than 98% are heparocytes).

101-

ROLE OF VEGF-FLT IN ANGIOGENESIS

Non parenchymal cells (Sinusoidal Endothelial Cells)

HGF

307

Parenchymal cells (Hepatocytes)

Met Receptor

___)

Flt receptor family

VEGF/VPF

FIG. 10. Model for two reciprocal communications in liver (Yamane et al., 1994).

i Liver

Sinusoid EC Kupffer etc.

__ HGF

Hepatocyte

VEGF

Arter. EC

Smooth MC

__ VEGF

(G)

Card. EC

?

Vent. Myocyte

VEGF

c3 Kidney

I,,(

Glomer. EC

3

4VEGF

Alveol. EC

__ ? VEGF

GI. Epithel. Mesang. cell Alveol. Epithel Alv Macroph 7

FIG. 1 1 . Possible reciprocal communication system through VEGF and other factors in various tissues. Abbreviations: NO, nitric oxide; ET, endothelin.

308

MASABUMI S H I B U Y A

tion of j t - 1 and KDRlflk-1 gene expression in vascular endothelial cells (Shweiki et al., 1993). 3 . Tumor Cells and Immortalized Cell Lines

In contrast to other receptor-type tyrosine kinases whose transcripts are easily detectable or overexpressed in tumor cells due to gene amplification, t h e j t - I mRNA was barely detectable in most of the human tumor cell lines examined so far (Shibuya et al., 1990). Further, no amplification nor rearrangement of the P t family genes in the chromosomes was reported in mammalian tumors. As an exception, P t - I was expressed in some choriocarcinoma cells and, very weakly, in K562 hematopoietic cells (Shibuya et al., 1990). Although KDRlflk-1 gene expression has not yet been studied extensively in tumors, Pt-4 mRNA was found in human erythroid leukemia (HEL), SK-NEP-1, and Y79 cells (Pajusola et ul., 1992). Melanoma cells, but not normal melanocytes, were found to carry VEGF receptors: theflt family member expressed in melanoma has not yet been identified (Gitay-Goren et al. 1993). T h e biological significance of these findings remains to be elucidated. As immortalized cells, human 293 and 293E1 cell lines obtained from embryo kidney by adenovirus transformation do express the @-I mRNA: 293E1 cells contain significant amounts of the f l t - I short message in addition to the full length mRNA. Although it is not clear whether the origin of these cells is renal endothelial cells, 293 and 293E1 cells might be useful for the characterization of the endogenous Flt-1 receptor (Shibuya e f ul., 1990). 4 . Regulatory Mechanism of Fit Gene Expression

Among the f l t gene family, characteristics of gene expression in flt-1 are as follows: (1) specificity to endothelial cells; (2) relatively high expression in endothelial cells until the middle stage of embryogenesis; and ( 3 ) high expression in endothelial cells during tumor angiogenesis compared to that of the surrounding normal vascular endothelial cells (Plate et al., 1992, 1993; Brown et al., 1993b). (1) and (2) hold true in the case of other flt-I-related genes, KDRlflk-1, andflt-4. On the basis of these findings, promoter-enhancer regions of theJ1t family in the genomic DNA should have an endothelial cell-specific enhancer and an additional enhancer that functions in the early stages of the embryo and in tumor angiogenesis. From this point of view, the r-&-I gene, which was shown to be highly specific to endothelial cells, seems to be a quite interesting candidate for the regulator of theflt gene family (Wernert et al., 1992).

309

ROLE OF VEGF-FLT IN ANGIOGENESIS

IV. Regulation of Tumor Growth by the Suppression of VEGF-Flt Receptor System A. SUPPRESSION OF TUMOR ANGIOGENESIS A considerable amount of evidence strongly suggests that (1) tumor angiogenesis is crucial for the rapid growth of solid tumors in uiuo and that (2) theVEGF-Flt receptor system plays an important role in the stimulation of tumor angiogenesis. Therefore, it is expected that the artificial suppression of VEGF-Flt receptors decreases the growth rate of tumors in vivo. Various steps in the VEGF-Flt receptor system could be possible candidates for suppressive control by chemicals, antibodies, and other peptides, as shown in Fig. 12. Kim et al., (1992, 1993) have shown that (1) monoclonal antibodies against the amino-terminal or carboxyl-terminal regions of VEGF molecule are able to suppress the VEGF activity and that (2) these monoclonal antibodies can block or decrease tumor growth in nude mice. A similar tumor-suppressive effect in viuo was also reported using polyclonal antibodies against VEGF (Kondo et al., 1993). Attempts to modulate the VEGF receptors or their functions have been reported. The soluble Flt-1 receptor, which is encoded in the natural short messages of thefit-I gene, was found to bind VEGF at a high affinity and could absorb the VEGF secreted from solid tumors, result-

&

Flt-1

t

KDR

9

5

@

Dominant negative R.

1. VEGF inhibitor (Ab etc.) 2. Absorption of VEGF 3. Dominant negative Receptor 4. Block of Ligand binding

5. Suppression of Ligand secretion 6. PlGF inhibitor (Ab etc.) 7. Inactivation of Extracellular domain 8.Inhibition of FItMDR tyrosine kinase 9. Block of Signaling to Specific Targets

FIG. 12. Suppression of tumor angiogenesis by controlling the VEGF-Flt receptor system.

310

MASABUMI SHIBUYA

ing in the suppression of endothelial cell growth and tumor angiogenesis (Kendall and Thomas, 1993). Further, a dominant-negative mutant (tyrosine kinase-negative mutant) of the KDR/Flk- 1 receptor efficiently blocked the signal transduction pathway of the VEGF-Flt receptor system when this truncated receptor molecule was highly expressed in endothelial cells (Millauer et d., 1994). Other steps such as the induction of VEGF gene expression in tumor cells and the possible endothelial cell-specific signal transduction machinery downstream from the Flt receptor family may also be possible good targets for tumor angiogenic blockers, although side effects due to the suppression of the VEGF-Flt system in normal tissues should be carefully controlled. Since FGFs are also known to contribute to tumor angiogenesis to some extent, the blockage of both the VEGF and FGF signaling pathways in parallel could have a better effect in some types of malignancies (Morrison, 1991).

B. CONTROL OF ASCITES FORMATION AND METASTASIS

VEGF was originally isolated from ascites-generating tumors. However, it is still not completely clear as to what extent the VEGF secreted from the tumor cells contributes to the formation of ascites. It seems most likely that the contribution of VEGF to ascites or pleural effusion is different in every patient with tumor. Thus, if one could evaluate the extent of VEGF’s contribution to the abnormal vascular permeability in each patient, blockage of the VEGF-Flt receptor system might be helpful in the treatment of patients associated with overexpression of VEGF. Although a close relationship between metastasis and the high level of VEGF in tumor tissues has not yet been demonstrated, one could expect that it is so, since the blood vessels are greatly involved in the metastatic process (Weidner et al., 1991; Senger et al., 1993).VEGF produced from tumor cells can stimulate permeability via activation o f cytoskeletal networks in the neighboring capillary endothelial cells: this kind of cytoskeletal disorganiLation of the vascular endothelial cells may also facilitate the migration of the tumor cells through the blood vessel wall. As described in this chapter, VEGF-FIt receptors appear to be the fundamental system for the development and maintenance of normal circular systems and most tissues, and this system is used in an abnormal way in pathological conditions, particularly in malignancy. Further insights into this novel ligand-receptor tyrosine kinase system in both basic science and tnedical applications hopefully will open a new field of “molecular biology of the endothelial cell” in biology.

ROLE OF VEGF-FLT IN ANGIOGENESIS

31 1

ACKNOWLEDGEMENTS I thank Dr. Lata Seetharam for reading the manuscript. I am also grateful to all the members in my laboratory for helpful discussions on the VEGF-Flt receptor system. REFERENCES

Adamis, A. P., Shima, D. T., Yeo, K.-T., Yeo, T.-K., Brown, L. F., Berse, B., DAmore, P. A., and Folkman, J. (1993).Biochem. Biophys. Res. Commun. 193, 631-638. Aiello, L. P., Avery, R. L., Arrigg, P. G., Keyt, B. A., Jarnpel, H. D., Shah, S. T., Pasquale, L. R., Thieme, H., Iwamoto, M. A., Park, J. E., Nguyen, H. V., Aiello, L. M., Ferrara, N., and King, G. L. (1994). New Eng1.J. Med. 331, 1480-1487. Aprelikova, O., Pajusola, K., Partanen, J., Armstrong, E., Alitalo, R., Bailey, S. K., McMahon, J., Wasmuth, J., Huebner, K., and Alitalo, K. (1992). Cancer Res. 52, 746748. Batzer, A. G., Rotin, D., Urena, J. M., Skolnik, E. Y., and Schlessinger,J. (1994). Mol. Cell. Bid. 14, 5 192-520 1. Berkman, R. A., Merrill, M. J., Reinhold, W. C., Monacci, W. T., Saxena, A., Clark, W. C., Robertson, J. T., Ali, I. U., and Oldfield, E. H. (19931.J. Clin. Invest. 91, 153-159. Berse, B., Brown, L. F., Van De Water, L., Dvorak, H. F., and Senger, D. R. (1992). Mol. Biol. Cell 3, 2 1 1-220. Bikfalvi, A., Sauzeau, C., Moukadiri, H., Maclouf, J., Busso, N., Bryckaert M., Plouet, J., and Tobelem, G. (1991).J. Cell. Physiol. 149, 50-59. Breier, G., Albrecht, U., Sterrer, S., and Risau, W. (1992). Development 114, 521-532. Brock, T. A., Dvorak, H. F., and Senger, D. R. (1991). Am. J. PathoE. 138, 213-221. Brown, L. F., Berse, B., Tognazzi, K., Manseau, E. J., Van De Water, L., Senger, D., Dvorak, H., and Rosen, S. (1992a). Kidney Int. 42, 1457-1461. Brown, L. F., Yeo, K.-T., Berse, B., Yeo, T.-K., Senger, D. R., Dvorak, H. F., and Van De Water, L. (1992b).J. Exp. Med. 176, 1375-1379. Brown, L. F., Berse, B., Jackman, R. W., Tognazzi, K., Manseau, E. J., Dvorak, H. F., and Senger, D. R. (1993a). Am. J. Pathol. 143, 1255-1262. Brown, L. F., Berse, B., Jackman, R. W., Tognazzi, K., Manseau, E. J., Senger, D. R., and Dvorak, H. F. (199313). Cancer Res. 53, 4727-4735. Charnock-Jones, D. S., Sharkey, A. M., Rajput-Williams, J., Burch, D., Schofield, J. P., Fountain, S. A., Boocock, C. A., and Smith, S. K. (1993). Bid. Reprod. 48, 1120-1128. Claffey, K. P., Wilkison, W. O., and Spiegelman, B. M. (1992).J. Bid. Chem. 267, 1631716322. Clauss, M., Gerlach, M., Gerlach, H., Brett, J., Wang, F., Familletti, P. C., Pan, Y.-C. E., Olander, J. V., Connolly, D. T., and Stern, D. (199O).J. Ex$. Med. 172, 1535-1545. Collins, P. D., Connolly, D. T., and Williams, T. J. (1993). Br. J. Pharmacol. 109, 195-199. Conn, G., Bayne, M. L., Soderman, D. S., Kwok, P.W., Sullivan, K. A., Palisi, T. M., Hope, D. A., and Thomas, K. A. (1990). Proc. Natl. Acud. Sci. L'SA 87, 2628-2632. Connolly, D. T. (1991).J. Cell Biochem 47, 219-223. Connolly, D. T., Heuvelman, D. M., Nelson, R., Olander, J. V., Eppley, B. L., Delfino, J. J., Siegel, N. R., Leimgruber, R. M., and Feder, J. (1989a).J. Clin. Invest. 84, 1470-1478. Connolly, D. T., Olander, J. V., Heuvelman, D., Nelson, R., Monsell, R., Siegel, N., Haymore, B. L., Leimgruber, R. M., and Feder, J. (1989b).J. Bid. Chem. 264,20017-20024. Cullinan-Bove, K., and Koos, R. D. (1993). Endoctinology 133, 829-837. De Vries, C., Escobedo,J. A., Ueno, H., Houck, K., Ferrara, N., and Williams, L. T. (1992). Science 255,989-991.

312

MASABUMI SHIBUYA

Dissen, G. A,, Lara, H. E., Fahrenbach, W. H., Costa, M. E., and Ojeda, S. R. (1994). Endocrinology 134, 1146-1 154. Dvorak, H. F., Sioussat, T. M., Brown, L. F., Berse, B., Nagy, J. A,, Sotrel, A., Manseau, E. J., Van De Water, L., and Senger, D. R. (1991). J . Exp. Med. 174, 1275-1278. Egan, S. E., Giddings, B. W., Brooks, M. W., Buday, L., Sizeland, A. M., and Weinberg, R. A. (1993). Nature 363, 45-51. Eichmann, A., Marcelle, C., Breant, C., and Le Douarin, N. M. (1993). Mechanism Dev. 42, 33-48.

Fant, W. J., Escobedo, J. A., Martin, G. A., Turck, C. W., del Rosario, M., McComick, F., and Williams, L. T. (1992). Cell 69, 413-423. Fava, R. A., Olsen, N. J., Spencer-Green, G., Yeo, K. T., Yeo, T. K., Berse, B., Jackman, R. W., Senger, D. R., Dvorak, H. F., and Brown, L. F. (1994).J. Exp. Med. 180,341-346. Ferrara, N., and Henzel, W. J. (1989). Biochem. Bophys. Res. Commun. 161, 851-858. Ferrara, N., Houck, K. A., Jakeman, L. B., Winer, J., and Leung, D. W. (1991a). J. Cell Biochpm 47,211-218. Ferrara, N., Leung, D. W., Cachianes, G., Winer, J., and Henzel, W. J. (1991b).Methods E n z y m l . 198, 39 1-405. Ferrara, N., Winer, J., and Burton, T. (1991c). Growth Factors 5, 141-148. Ferrara, N., Houck, K., Jakeman, L., and Leung, D. W. (1992). Endocnnol Rev. 13, 18-32. Ferrara, N., Winer, J., Burton, T., Rowland, A., Siegel, M., Phillips, H . S., Terrell, T., Keller, G. A.. and Levinson, A. D. (1993). J. Clin. Invest. 91, 160-170. Finnerty, H., Kelleher, K., Morris, G. E., Bean, K., Merberg, D. M., Kriz, R., Morris, J. C., Sookdeo, H., Turner, K. J., and Wood, C. R. (1993). Ontogene 8,2293-2298. Folkman, J. (1990). J. Natl. CameiInst. 82, 4-6. Galland, F., Karamysheva, A,, Pebusque, M.-J., Borg, J.-P., Rottapel, R., Dubreuil, P., Rosnet, 0..and Birnbaum, D. (1993). Oncogene 8, 1233-1240. Garrido, C., Saule, S., and Gospodarowicz, D. (1993). Growth Factors 8, 109-1 17. Gitay-Goren, H., Soker, S., Vlodavsky, I., and Neufeld, G. (1992).j. Biol. Chem. 267,60936098.

Gitay-Goren, H., Halaban, R., and Neufeld, G. (1993). Biochem. Biophys. Res. Commun. 190, 702-709.

Goldberg, M. A., and Schneider, T. J. (1994).j. Biol. Chem. 269, 4355-4359. Goldman, C. K., Kim, J., Wong, W. L., King, V., Brock, T., and Gillespie, G. Y.(1993). Mol. Biol. Cell 4, 121-133. Gospodarowicz, D., Abraham, J. A., and Schilling, J. (1989). Proc. Natl. Acud. Sci. USA 86, 731 1-7315.

Goto, F., Goto, K., Weindel, K., and Folkman, J.(1993). Lab. Invest. 69, 508-5 17. C h o h , N., Tojo, A., Muroya, K., Hashimoto, Y., Hattori, S., Nakamura, S., Takenawa, T., Yazaki. Y.,and Shibuya, M. (1994). Proc. Nati. Acod. Sci. USA 91, 167-171. Harada, S., Nagy, J. A,, Sullivan, K. A., Thomas, K. A., Endo, N., Rodan, G. A., and Rodan, S. B. (1994). J. Clin. Invest. 93, 2490-2496. Hauser, S., and Weich, H. A. (1993). Growth Factors 9, 259-268. Houck, K. A., Ferrara, N., Winer, J., Cachianes, G., Li, B., and Leung, D. W. (1991). Mol. Endocrinol. 5, 1806-1814. Houck, K. A,, Leung, D. W., Rowland, A. M., Winer, J., and Ferrara, N. (1992). J . Biol. Chem. 967, 26031-26037. Humphrey, P. A,, Wong, A. J., Vogelstein, B., Zalutsky, M. R., Fuller, G. N., Archer, G. E., Friedman, H. S., Kwatra, M. M., Bigner, S. H., and Bigner, D. D. (1990). Proc. Natl. Acud. Sci. USA 87,4207-42 1 1.

ROLE OF VEGF-FLT IN ANGIOGENESIS

313

Iijima, K., Yoshikawa, N., Connolly, D. T., and Nakamura, H. (1993).Kadney Int. 44, 959966. Jakeman, L. B., Winer, J., Bennett, G. L., Altar, C. A,, and Ferrara, N. (1992).J.Clin. Invest. 89,244-253. Jakeman, L. B., Armanini, M., Phillips, H. S., and Ferrara, N. (1993). Endocrinology 133, 848-859. Kaipainen, A., Korhonen, J., Pajusola, K., Aprelikova, O., Persico, M. G., Terman, B. I., and Alitalo, K. (1993).J. Exp. Med. 178, 2077-2088. Keck, P. J., Hauser, S. D., Krivi, G., Sanzo, K., Warren, T., Feder, J., and Connolly, D. T. ( 1989). Science 246, 1309- 13 12. Kendall, R. L., and Thomas, K. A. (1993). Proc. Natl. Acad. Sci. USA, 90, 10705-10709. Kendall, R. L., Wang, G., DiSalvo, J., and Thomas, K. A. (1994). B i o c h a . Biophys. Res. Commun. 201,326-330. Kieser, A., Weich, H. A,, Brandner, G., Marme, D., and Kolch, W. (1994).Oncogme 9,963969. Kim, K. J., Li, B., Houck, K., Winer, J., and Ferrara, N. (1992). Growth Factors 7, 53-64. Kim, K. J., Li, B., Winer, J., Armanini, M., Gillett, N., Phillips, H. S., and Ferrara, N. ( 1993). Nature 362, 84 1-844. Koch, A. E., Harlow, L. A., Haines, G. K., Amento, E. P., Umemori, E. N., Wong, W. L., Pope, R. M., and Ferrara, N. (1994).J. Immunol. 152,4149-4156. Kondo, S., Asano, M., and Suzuki, H. (1993). Bzocha. Biophys. Res. Commun. 194, 12341241. Ku, D. D., Zaleski, J. K., Liu, S., and Brock, T. A. (1993).Am.J. Physiol. 265 (2 Pt 2), H586592. Ladoux, A., and Frelin, C. (1993). Biochem. Biojhys. Res. Cmnmun. 195, 1005-1010. Leung, D. W., Cachianes, G., Kuang, W.-J., Goeddel, D. V., and Ferrara, N. (1989). Science 246, 1306- 1309. Libermann, T. A., Nusbaum, H. R., Razon, N., Kris, R., Lax, I., Soreq, H., Whittle, N., Waterfield, M. D., Ullrich, A,, and Schlessinger, J. (1985). Nature 313, 144-147. Lyttle, D. J., Fraser, K. M., Fleming, S. B., Mercer, A. A., and Robinson, A. J. (1994).J. Virol. 68, 84-92. Maglione, D., Guerriero, V., Viglietto, G., Delli-Bovi, P., and Persico, M. G. (1991). Proc. Natl. Acad. Sci. USA 88,9267-9271. Maglione, D., Guerriero, V., Rambaldi, M., Russo, G., and Persico, M. G. (1993a).Growth Fators 8, 141-152. Maglione, D., Guerriero, V., Viglietto, G., Ferraro, M. G., Aprelikova, O., Alitalo, K., Vecchio, S. D., Lei, K.-J., Chou, J. Y., and Persico, M. G. (1993b).Oncogene 8,925-93 1 . Matsushime, H., Yoshida, M. C., Sasaki, M., and Shibuya, M. (1987).Jpn.J. CuncerRes. 78, 655-66 1. Matthews, W., Jordan, C. T., Gavin, M., Jenkins, N. A., Copeland, N. G., and Lemischka, 1. R. (1991). Proc. Natl. Acad. Sci. USA 88, 9026-9030. McClure, N., Healy, D. L., Rogers, P. A. W., Sullivan, J., Beaton, L., Haning, R. V., Jr., Connolly, D. T., and Robertson, D. M. (1994). Lancet 344, 235-236. Midy, V., and Plouet, J. (1994). Biochem. Bisphys. Res. Commun. 199, 380-386. Millauer, B., Wizigmann-Voos, S., Schnurch, H., Martinez, R., Moller, N. P. H., Risau, W., and Ullrich, A. (1993). Cell 74, 835-846. Millauer, B., Shawver, L. K., Plate, K. H., Risau, W., and Ullrich, A. (1994). Nature 367, 576-579. Miller, J. W., Adamis, A. P., Shima, D. T., DAmore, P. A., Moulton, R. S., OReilly, M. S.,

3 14

MASABUMI SHIBUYA

Folkman, J., Dvorak, H . F., Brown, L. F.. Berse, B., Yeo, T.-K., and Yeo, K.-T. (1994). Am. J. Pathol. 145, 574-581. Morii, K., Tanaka, R., Washiyama, K., Kumanishi, T., and Kuwano, R. (1993). Biochem. Biophy. Res. Cornrnun. 194, 749-755. Morrison, R. S. (1991).J. B i d . Chum. 266, 728-734. Myoken, Y., Kayada, Y., Okamoto, T., Kan, M., Sato, G. H., and Sato, J. D. (1991). Proc. NaTad. ilcad. Scz. USA 88, 5819-1823. Olander, J . V., Cunnolly, D. T., and DeLarco, J . E. (1991). Biochem. Biophys. Res. Commun. 175,68-76. Olson, T. A., Mohanraj, D., Carson, L. F., and Ramakrishnan, S. (1994). Cancer Res. 54, 276-280. Pajusola, K., Aprelikova, O., Korhonen, J., Kaipainen, A., Pertovaara, L., Alitalo, R., a n d Alitalo, K. (1992). Cancer Res. 52, 5738-5743. Pajusola, K., Xprelikova, O., Armstrong, E., Morris, S., and Alitalo, K. (1993). Oncogene, 8, 293 1-2937. Pajusola, K., Aprelikova, 0.. Pelicci, G., Weich, H., Claesson-Welsh, L., and Alitalo, K. (1994). Oncogme 9, 3545-3555. Park, J. E., Keller, G. A., and Ferrara, N. (1993). Mol. B i d . Cetl 4, 1317-1326. Pekala, P., Marlow, hf., Heuvelman, D., and Connolly, D. (1990).J. Biol. Chem. 265, 1805118054. Pepper, M. S., Wasi, S., Ferrara, N., Orci, L., and Montesano, R. (1994). Exp. Crll. Res. 210, 298-305. Peretz, D., Gitay-Goren, H., Safran, M., Kimmel, N., Gospodarowicz, D., and Neufeld, G. (1992). Biochem. Biophjs. Res. Cornmun,. 182, 1340-1347. Pertovaara, L., Kaipainen, A., Mustonen, T., Orpana, A,, Ferrara, N., Saksela, O., and Alitalo, K. (1 994).J . Biol. Chem. 269, 627 1-6274. Peters, K. G., De Vries, C., and Williams, L. T. (1993). Proc. ,\‘ad. Aced. Scz. USA 90, 89158919. Phillips, H. S.. Hains, J . , Leung, D. W., and Ferrara, N. (1990). Eiidocirnology 127, 965967. Plate, K. H., Breier, (;., Weich. H. A . , and Risau, W. (1992). ‘Valialure 359, 845-848. Plate, K. H.. Breier, (;., Millauer, B., Ullrich, A,, and Risau, W. (1993). Cancer Re.\. 53, 5822-5825. Plouet. J . , and hloukadiri, H. (1990).J. B i d . Chem. 265, 22071-22074. Plouet, J . , Schilling, J., and Gospodarowicz, D. (1989). E M B O J . 8, 3801-3806. Quinn, ?I P., Peters. K. G., De Vries, C., Ferrara, N., and Williams, L. T. (1993). Pror. Natl. Acad. Sci. G S A 90, 7533-7537. Ravindranath, N.. Little-Ihrig, L., Phillips, H. S., Ferrara, N., and Zeleznik, A. J. (1992). Endoctinolorn 131, 254-260. Risau, W.. Sariola, H., Zerwes, I-1.-G., Sasse, J., Ekblom, P., Keniler, K.,and Doetschman, T. (1988).Dezjeloptnenl 102, 47 1-478. Rosenthal, R. A., Megyesi, J. F., Henzel, W. J.. Ferrara, N., and Folkman, J. (1990). Growth Fn*-/nr.c 4, 53-.59. Rosnet, O., Stephenson, D., Mattei, hf.-G., Marchetto, S., Shibuya, M., Chapman, V. M., and Birnbaum, D. (1993). Oncogene 8, 175-179. Sarzani, R., Amaldi, G., De Pirro, R., hforetti, P., Schiaffino, S., and Rappelli, A. (1992). Biochtm. BiSp/zy.s. Re.s. Comniun. 186, 706-7 14. Satoh, H.. Yoshida, bf. C., Matsushime, H., Shibuya, M.,and Sasdki, M. (1987). ]pi.J . Ganrvr Res. 78. 772-77.3.

ROLE OF VEGF-FLT IN ANGIOGENESIS

315

Schlessinger, J., and Ullrich, A. (1992). Neuron 9, 383-391. Seetharam, L., Gotoh, N., Maru, Y., Neufeld, G., Yamaguchi, S., and Shibuya, M. (1995). Oncogene 10, 135-147. Senger, D. R., Galli, S. J., Dvorak, A. M., Perruzzi, C. A., Harvey, V. S., and Dvorak, H. F. (1983). Science 219, 983-985. Senger, D. R., Connolly, D. T., Van De Water, L., Feder, J., and Dvorak, H. F. (1990).Cancer Res. 50, 1774-1778. Senger, D. R., Van d e Water, L., Brown, L. F., Nagy, J. A., Yeo, K. T., Yeo, T. K., Berse, B., Jackman, R. W., Dvorak, A. M., and Dvorak, H. F. (1993). Cancer Metastasis Rev. 12, 303-324. Sharkey, A. M., Charnock-Jones, D. S., Boocock, C. A,, Brown, K. D., and Smith, S. K. ( 1993).J. Reprod. F e d . 99, 609-6 15. Shen, H., Clauss, M., Ryan, J., Schmidt, A.-M., Tijburg, P., Borden, L., Connolly, D., Stern, D., and Kao, J. (1993).Blood 81, 2767-2773. Shibuya, M., Yamaguchi, S., Yamane, A,, Ikeda, T., Tojo, A., Matsushime, H., and Sato, M. (1990). Oncogene 5, 519-524. Shifren, J. L., Doldi, N., Ferrara, N., Mesiano, S., and Jaffe, R. B. (1994).J.Clin. Endocn'nol. Metab. 79, 316-322. Shweiki, D., Itin, A,, Soffer, D., and Keshet, E. (1992). Nature 359, 843-845. Shweiki, D., Itin, A,, Neufeld, G., Gitay-Goren, H., and Keshet, E. (1993).J.Clin. Invest. 91, 2235-2243. Sioussat, T. M., Dvorak, H. F., Brock, T. A,, and Senger, D. R. (1993). Arch. Biochem. Biophys. 301, 15-20. Takahashi, J. A., Mori, H., Fukumoto, M., Igarashi, K., Jaye, M., Oda, Y., Kikuchi, H., and Hatanaka, M. (1990). Proc. Natl. Acad. Sci. USA 87, 5710-5714. Takahashi, A., Sasaki, H., Kim, S. J., Tobisu, K., Kakizoe, T., Tsukamoto, T., Kumamoto, Y., Sugimura, T., and Terada, M. (1994). Cancer Res. 54, 4233-4237. Terman, B. I., Carrion, M. E., Kovacs, E., Rasmussen, B. A., Eddy, R. L., and Shows, T. B. (1991). Oncogene 6, 1677-1683. Terman, B. I., Dougher-Vermazen, M., Carrion, M. E., Dimitrov, D., Armellino, D. C., Gospodarowicz, D., and Bohlen, P. (1992). Biochem. Biophys. Res. Commun. 187, 15791586. Tessler, S., Rockwell, P., Hicklin, D., Cohen, T., Levi, B. Z., Witte, L., Lemischka, I. R., and Neufeld, G. (1994).J.Biol. Chem. 269, 12456-12461. Tischer, E., Mitchell, R., Hartman, T., Silva, M., Gospodarowicz, D., Fiddes, J. C., and Abraham, J. A. (1991).J.B i d . Chem. 256, 11947-1 1954. Uchida, K., Uchida, S., Nitta, K., Yumura, W., Marumo, F., and Nihei, H. (1994). Am. J . Physiol. 266 ( 1 Pt 2), F81-88. Unemori, E. N., Ferrara, N., Bauer, E. A., and Amento, E. P. (1992).J. Cell. Physiol. 153, 557-562. Vaisman, N., Gospodarowicz, D., and Neufeld, G. (199O).J.Biol. Chem. 265,19461-19466. Valius, M., and Kazlauskas, A. (1993). Cell 73, 321-334. Waltenberger, J., Claesson-Welsh, L., Siegbahn, A,, Shibuya, M., and Heldin, C.-H. (1994). J. Biol. Chem. 269, 26988-26995. Wang, R., Clark, R., and Bautch, V. L. (1992).Development 114, 303-316. Weidner, N., Sernple, J. P., Welch, W. R., and Folkman, J. (1991). New Engl. J. Med. 324, 1-8.

Weindel, K., Marme, D., and Weich, H. A. (1992). Biochem. Bzophys. Res. Commun. 183, 1167-1 174.

3 16

MASABUMI SHIBUYA

Wernert, N., Raes, M. B., Lassalle, P., Dehouck, M. P., Gosselin, B., Vandenbunder, B., and Stehelin, D. (1992). A m . ] . Pathol. 140, 119-127. Wilting, J., Christ, B., and Weich, H. A. (1992). Anal. Embryol. 186, 251-257. Wilting, J., Christ, B., Bokeloh. M., and Weich, H. A. (1993). Cell TissueRes. 274, 163-172. Wrana, J. L., Attisano, L., Wieser, R., Ventura, F., and Massague, J. (1994). Nature 370, 34 1-347. Yamaguchi, T. P., Dumont, D. J., Conlon, R. A., Breitman, M. L., and Rossant, J. (1993). Development 118, 489-498. Yamane, A,, Seethardm, L., Yamaguchi, S., Gotoh, N., Takahashi, T., Neufeld, G., and Shihuya, M. (1994). Oncogene 9, 2683-2690. Yamazaki, H., Fukui, Y., Ueyama, Y., Tamaoki, N., Kawamoto, T., Taniguchi, S., a n d Shibuya, M. (1988). Mol. Cell. B i d . 8, 1816-1820. Yamazaki, H., Ohba, Y., Tamaoki, N., a n d Shibuya, M. (1990).Jpn.]. Cancer Res. 81, 773779. Yamazaki, K., Sato, K., Shizume, K., Kanaji, Y., Demura, H., Ito, Y., Obara, T., Sawano, A., Ishii, Y., Yamaguchi, S., and Shihuya, M. (1994). The fQth Asia-Oceania Congress of Endocrinology (Abstracts). Yan, Z., Weich, H. A . , Bernart, W., Breckwoldt, M., and Neulen, J. (1993).J. Clzn. Endocrinol. Metab. 77, 1723- 1725. Yeo, T.-K., Senger, D. R., Dvorak, H. F., Freter, L., and Yeo, K.-T. (1991). Bzochem. Biophys. Rex Commun. 179, 1568-1575.

INDEX

A Acute lymphoblastic leukemia childhood, 36-38 remissions, 2 Acute lymphoid leukemia, chromosomal translocations, 27-38 Acute myeloid leukemia, EVIl expression pattern, 38-40 Acute promyelocytic leukemia, transcription factor fusion, 39-40 Age, dependent incidence of cancer, 5-7 AIDS-DLCL, 116-142 AIDS-NHL B-cell clonal expansion, 118-120 EBV pathogenic role, 126-130 epidemiology, 113-1 14 HIV role, 130-131 AIDS-PCNSL, 115-142 AIDS-related Kaposi’s sarcoma angiogenesis, SF role, 268 epidemiology, 113-1 14 AIDS-SNCCL, 116-141 ALL, see Acute lymphoblastic leukemia Allelic losses, see also Loss of heterozygosity HLA, from tumor cells, 172-173 a 3 domain, contact with &-microglobulin, 161-162 Amino acid structure, VEGF, 284-286 Amplification c-myc, 12 EGFR gene, 292 Anaplastic large cell lymphomas, AIDSrelated, systemic, 116-1 17 Angiogenesis in human cancers, 266-267

317

in vavo assays, 263-266 regulation by VEGF-Flt receptors, 281282 role of SF, 271-275 tumor macrophage role, 269-270 PDGF role, 295-296 suppression, 309-3 10 Antibodies EBV-specific responses, 237-239 recombinant chimeric, 275 Antigens chronic stimulation, role in AIDSSNCCL pathogenesis, 123-124 HLA class I expression in normal tissue, 157-166 lack of expression, 178- 180 LFA, down-regulation, 135 lymphocyte-determined membrane, 239 SV40 large T, transgenic mice expressing, 99-100 tumor, and HLA, 170-171 tumor-associated transplantation, polymorphism, 170- 171 Antioncogenes, see also Tumor suppressor genes interaction with tumor antigens, 18 p53, 272 RBI, 12-14 Antiretroviral therapy, 114, 122 Ascites formation, role of VEGF, 310 A-T hook motif, in human acute leukemia, 26-38 Autophosphorylation, Flt family, 302303 Avian myeloblastosis virus, effects resembling Wilms’ disease, 2-3

318

INDEX

C Back-crossing, mouse strains, 95-99 BamHI fragments EBV genome, 200-244 N fragment, LMP-I coding region in, 214-2 18 Basic helix-loop-helix motif, in human acute leukemia, 26-33 B-cell acute leukemia, see Burkitt’s lymphoma B cells chronic antigenic stimulation, 122124 clonal expansion in AIDS-NHL, 118120 differentiation, effect of c-MYC protein expression, 135- 136 expansions, EBV-infected, 126-130 latently infected with EBV, 200-204 bcl-2, transcription, inhibition by p53, 272 BCL-6, as prognostic marker, 136 Biological functions, Flt family, 301302 Biology EBV infections, 241-246 role of VEGF in normal tissues, 290 Blood vessels, permeability, in cancer progression, 282 Bone marrow transplant, and EBV transmission, 198-199, 235-236, 244 Break points on chromosomes 8 and 14, 134 in human acute leukemia, 28-38 in human sarcomas, 40-43 T-cell translocations, reverse genetics at, 50 Breast cancer abnormal karyotypes, 68-7 1 RER+ phenotype, 74-75 SF and tumor angiogenesis as prognostic indicators. 267 Burkitt’s lymphoma chromosomal translocations, 1 1, 27-3 1 EBV’, 198, 202-204 HIV+ and H I V - , 21 1-213 sporadic, similar to AIDS-SNCCL, 134

Cancer, see also Sarcomas; Tumors age-specific incidence, 5-7 hereditary, 7-9 human angiogenesis, 266-267 in children, 1-2 progression in oivo, role of blood ves. sels, 282 Carcinogenesis, genetic events, 16-1 8 Carcinogens induction of lung tumors in mice, 8791 and somatic mutations, 4-7 Carcinomas, see also Sarcomas; Tumors SF and c-met up- and down-regulation, 267 Carrier status, EBV, 231-234 Case-control studies, lung tumors in humans, 102-104 CD4 counts in AIDS-NHL, 121-122 decreased, and lymphoma development, 139-142 in EBV+ AIDS-NHL, 127-130 Cell fate, role of oncogenic transcription factors, 48-50 Cell lines immortalized, and tumor cells, fit-I mRNA, 308 Jijoye, Burkitt’s lymphoma-derived, 220-221 lymphoblastoid EBV ebnotypes in, 23 1-234 with EBV latency 111, 202-210 EBV-transformed, 226-23 1 NIH3T3, Flt-overexpressing, 303 c-fos, relation to MHC class I expression, 178 Chaperones, heat shock proteins as, 163164 Children, cancer in, 1-2 Chimeras AML1-ETO protein, 38-39 E2A-HLF, 48-50 EZA-PBXI, 32-34 Chromosomes, see also Philadelphia chromosome

319

INDEX

alteration pathways in human epithelial cancers, 59-78 LOH on chromosome 18, 61-62 in murine and human lung tumors, 93-95 8q24 translocations, 132-134 1lq23 abnormalities, 36-38 rearranged, distribution in relation to ploidy, 69-7 1 17, H 2 complex, 98-99 6 Kras2 locus, 97- 102 Pa~l~OCUS, 97-102, 105 13, esterase D gene on, 13-14 X, MACE-1 on, 171 Chronic antigen stimulation, role in AIDS-SNCCL pathogenesis, 123124 Chronic myelocytic leukemia abnormal karyotype, 5 genetic events, 16 reciprocal translocation, 10 Cis-acting regulatory sequences, in class I expression, 157- 159 Classification nuclear proteins, 26-27 pathological, AIDS-NHL, 115-1 16 Class I regulatory element, role in determining class I expression, 159-160 Clinical applications, scatter factor, 274275 Clinicopathologic spectrum, AIDS-related lymphomas, 115-1 18 c-met, RNA expression, 273-274 c-met receptor, SF-activated, 258 CML, see Chronic myelocytic leukemia C-MYC rearrangements in AIDS-NHL, 120 role in AIDS-related lymphomas, 132136 c-my, mediated MHC class I downmodulation, 177-178 Colorectal adenocarcinoma, cytogenetics, 61-68 Comparative mapping, human and murine chromosomal regions, 104 Cornea, neovascularization, in vivo assay, 264-266

CTL alp CD8+, 166-168 EBV-specific reactivities, 239-24 1 epitopes HLA-A 1 1-restricted, 240-244 peptide candidates as, 167-168 Cytogenetics, colorectal adenocarcinoma, 61-68 Cytokines, see also spec@ cytokines deregulation in AIDS-related lyrnphomagenesis, 124- 126 regulation of SF production, 270-271

D Deletions 30-base pair, and single-base mutations, 217-2 18 long arm of chromosome 6, 138-139 in 17p arm in monosomic-type tumors, 67-68 in 51q arm in familial adenomatous polyps, 65 unmasking of recessive mutations, 5961 Development AIDS-related lymphomas contributing factors, 120- 126 pathogenetic pathways, 139-142 normal mammalian, role of oncogenic transcription factors, 46-48 Developmental model, master transcription factors, 25-51 Diffuse large cell lymphomas, AIDSrelated absence of BCL-2 rearrangements, 137 disrupted immunosurveillance, 120-122 role of c-MYC, 132-133 systemic, 116-1 17, 127 Dimerization domains, homologies with DNA-binding domains, 43-46 DLCL, see Diffuse large cell lymphomas DNA and EBV typing, 204-226 hybridization studies, 10- 12 repair, p53 signal for, 15-16 viral, effect on class I molecules, 181182

320

INDEX

DNA-binding domains, homologies with dimerization domains, 43-46 DNA-binding factors, class 1 promoter, 159- 160 DNA ploidy, in breast and colorectal cancers, 70-71 DNA tumor viruses, activities of RBI and TP53, 14-16 Down-regulation HLA alleles, I56 LFA antigens, 135 and up-regulation, SF and c-met, 267 Drosophiln, developmental regulatory proteins, 43-50

E EBER-1 , and EBER-2. transcription, 202204 EBNA-I, in EBV-transformed cells, 227 EBNA-2 in AIDS-related lymphomas, 141-142 defective EBV genotypes, 220-224 encoded by EBV types A and B, 227228 expression patterns in AIDS-NHL, 129- 130 PCR assay, 21 1 EBNA-3, in ebnotyping, 228-230 EBNA-4 in ebnotyping, 228-230 type A, containing HLA-A 1 I-restricted CTL epitopes, 240-241 EBNA-5, consistent with BamHI W exon coding capacity, 230 EBNA-6 in EBV-transformed cells, 228-230 PCR assay, 21 1 EBNAs mRNA expression, 200-204 polymorphisms, 236-237 Ebnotyping, Epstein-Sarr virus, 226237 EBV, s e p Epstein-Barr virus EC, .we Endotheld cells Embryogenesis Flt gene family expression, 304-308 VEGF mRNA levels during, 289290

Endometrial cancer cytogenetic data, 73-74 microsatellite instability, 76-77 Endoreduplication in breast cancer, 69-70 in non-small-cell lung cancer, 7 1-73 Endothelial cells derived tumor cells, VEGF expression, 293 proliferation and tubular formation, 286-288 vascular, as SF producer cell and target cell, 259-261 Enhancer A, regulatory element, 158-159 Epidemiology AIDS-related lymphomas, 113-1 14 EBV infection, 197-246 Epithelial cancers, human, chromosome alteration pathways, 59-78 Epithelial tumors EBV+, 198-199 microsatellite instability, 74-77 Epstein-Barr virus carrier status, 231-234 ebnotyping, 226-237 genome and gene expression, 200-204 genotypes defective, 220-224, 244-245 differential recognition by immune system, 237-241 genotypes and ebnotypes, sorting, 24 1243 genotyping, 204-226 pathogenesis, 244-246 role in AIDS-related lymphomagenesis, 126-130 transmission, study with RFLP, 224 types A and B, differential detection, 205-2 13 Esrerase D, and recessive hypothesis, 13 Evolution monosomic, breast cancers, 75 phylogenetic, fm/f?t supergene family, 297-298 viral, under immunological pressure, 243-244 Ewing’s sarcoma, t( 11;22) translocation, 40-43 EWS domain, fusions in human sarcomas, 42-43

32 I

INDEX

F Familial adenomatous polyps, and tumor development, 65 Familial studies, lung tumors in humans, 102- 104 Fibroblast growth factor angiogenic activity, 282-283 basic, expression in glioblastoma multiform, 292 induced angiogenesis, 265-266 Flt receptor family, for VEGF, 296-308 biological functions, 301-302 signal transduction, 302-304 structural characteristics, 297-30 1 Fms/Kit/PDGFR, 5-Ig-type receptors, 297-299 Fusion genes BCR-ABL, 32-33 E2A-PBXl, 32-34 E WS- W T l , 42-43

G Gap proteins, regulation of pair-rule gene expression, 45-46 Gene expression and EBV genome, 200-204 Flt family, 304-308 in human and murine lung cancers, 91-93 pair-rule, regulation by gap proteins, 45-46 VEGF, 289-294 Genes, see a h Fusion genes; Tumor suppressor genes BCL-6, 136 BNLFl, 205, 214-218 c-my, 12 DCC, 17 jlt-1

and jlt-4, 300-30 1 and KDR, 297 HLA, 184 HLF, 34-36 H O X l I , 32-33,46-47 human cancer, 7-12 interstitial collagenase, 288 K r a 2 , 86-91

lung tumor susceptibility, candidate, 100-102 MLL, 36-38 mutations, in murine lung tumors, 8691 MYC and TCR, 27-31 p53,91 PAX, 47 peptide transported, abnormalities, 180 rm, 11-12 RBI and TP53, 14-18 temperature-sensitive, Rous sarcoma virus, 4 TP53, 67-68,75 transcriptional control, alteration, 26-43 Genetic events, in carcinogenesis, 16- 18 Genetic lesions in AIDS-NHL, 120 preceded or followed by chronic antigen stimulation, 124 role in AIDS-related lymphomas, 132139 Genetic linkage studies, in murine lung tumorigenesis, 95-99 Genetics human lung tumors, 102-104 murine lung tumors, 83- 106 Genomic alterations, pathways, and tumor type, 67-68 Genomic heterogeneity, LMP- 1 coding region in BamHI N fragment, 214-218 Genotyping, Epstein-Barr virus, 204-226 Geographic distribution C/D and Flf polymorphisms, 219-220 EBV genotypes, 210-213, 224-226 Glioblastoma multiform, expression of VEGF mRNA, 291-292 Groove, class I, peptide-binding, 164-165, 167

H H3-8, RFLP, 14 H2 complex, on chromosome 17, 98-99 Heavy chains HLA class I, 161 MHC class I, abnormalities, 176-178 Hematological tumors, and EBV occurrence, 211-212

322

INDEX

Hepatocyte growth factor, and c-Met receptor, 306 Hereditary nonpolyposis colorectal cancers and absence of breast cancer, 75 and tumor development, 66-67 Histology, murine lung tumors, 85-86 HIV, see Human immunodeficiency virus H-2Kb, abrogated expression, 176 H-2Kk, private specificity loss, 156- 157 HLA class I expression, alteration in human tumors, 171- 183 molecular assembly, 161-165 and tumor antigens, 170- 171 Iiodgkin’s disease, lesions, containing EBV types A and B, 212-213 Human cancer angiogenesis, 266-267 in children, 1-2 lung tumors, genetics, 102-104 Human cytomegalovirus, effect on MHC class I expression, 181-182 Human immunodeficiency virus association with Hodgkin’s disease, 114 with EBV, 210-213 etiologic agent of AIDS-NHL, 130-131 Hybridization, DNA, 10-12 Hypoxia, induced VEGF mRNA, 2932 94

I Immune system, differential recognition of EBV genotypes, 237-241 Immunohistological techniques, insensitivity, 165-166 I mm unoselect ion MHC-driven, 156- 157 randomly produced tumor cell clones, 183 Inimunosurveillance disrupted, and development of AIDSrelated lymphoma, 120-122 in EBV pathogenesis, 245-246 Immunotherapy, HLA active, 184 Infection, EBV, molecular epidemiology, 197-246

infectious mononucleosis EBV type A in, 212-213 with productive EBV infection, 198199 Inherited predisposition to lung cancer in humans, 103-105 to lung tumorigenesis in mice, 84-95 Injurin, SF-inducing, 271 Interferon-alp, induced gene transcription, 160-161 Interferon-y, in p91 activation, 160-161 Interleukin-1, stimulation of SF production, 270-271 Interleukin-6, direct stimulation by HIV, 125-126 Interleukin-10, role in AIDS-related lymphoproliferations, 125- 126

K Kaposi’s sarcoma AIDS-related angiogenesis, role of SF, 268 epidemiology, 1 13- 114 VEGF mRNA levels, 293 Karyotype abnormal in breast cancer, 68-71 in CML, 5 alterations, in tumor subset, 63-64 normal in hereditary nonpolyposis colorectal cancer, 66 in tumor subset, 64 KDRIFlk-1 receptors, VEGF-binding, 301-302 Keratinocytes, near wound lesion, expression of VEGF mRNA, 291 KRAS2 activation, in pathogenesis of lung adenocarcinoma, 86-9 1 Kras2 locus, on chromosome 6, 97-102 Kringle domains, in SF truncated form, 274-275

L Latency patterns, EBV, 129-130, 20020 1, 245-246

INDEX

LCL, see Lymphoblastoid cell lines Leukemias acute lymphoblastic, remissions, 2 acute lymphoid, chromosomal translocations, 27-38 acute myeloid, 38-40 B-cell acute, see Burkitt’s lymphoma E2A-HLF-associated, 34-36 Leukoplakia, oral hairy EBV+ biopsies, 221-224 PCR analysis, 2 17 Li-Fraumeni syndrome, TP53 mutant in, 17-18 LIM domains, transcriptional regulatory proteins, 31-32 LMP- 1 in AIDS-related lymphomas, 141-142 cytoplasmic domains, 245 expression patterns in AIDS-NHL, 129- 130 genomic heterogeneity in BamHI N fragment, 214-218 and LMP-2A and -2B expressed during EBV latency, 200204 transcription, 230-23 1 LOD score, for lung tumors, 95-98 LOH, see Loss of heterozygosity Losses, see also Allelic losses MHC class I, mechanisms, 182-183 total, locus, and allelic, in HLA phenotype, 171-173 XhoI restriction site, 215-218, 241243 Loss of heterozygosity in murine and human lung tumors, 9395 for 17p and 18q arms, 61-62 Lung cancer microsatellite instability, 76 non-small-cell distribution of rearrangements, 7173 familial studies, 103-104 p53 inactivation in, 91 Lung tumors genetics human, 102-104 murine, 83-106 susceptibility genes, 100- I02

323

Lymphadenopathy, persistent generalized in early AIDS-NHL, 118-120 preceding AIDS-SNCCL, 139-142 presence of EBV in context of, 128129 Lymphoblastoid cell lines EBV ebnotypes in, 231-234 with EBV latency 111, 202-210 EBV-transformed, 226-23 1 Lymphocytic function antigen-I, downregulation, 135 Lymphomagenesis, AIDS-related cytokine role, 124-126 viral infection role, 126-132 Lymphomas, see also specijic types AIDS-related clinicopathologic spectrum, 115-1 18 development, 120-126, 139-142 epidemiology, 1 13- 1 14 role of genetic lesions, 132-139 peripheral T-cell 1 1-aa repeats, 2 17 EBV+, 198-199 EBV genotypes, 204-205

mAb, see Monoclonal antibodies Macrophages, role in tumor angiogenesis, 269-270 MACE-1, on chromosome X, 17 1 Major histocompatibility complex class I expression alteration in tumor cells, 175-182 gene regulation, 157-161 viral infection effects, 180-182 role in T and NK cell recognition, 166169 MAP kinase, cascade activation by Flt family, 303-304 Mapping, comparative, human and murine chromosomal regions, 104 Mathematical model, number of events in cancer, 5-7 Matrigel ECs plated onto, 260-261 in mouse assay of angiogenesis, 263266

524

INDEX

Messenger RNA EBNA, 200-204 flt-1, 305-308 surfactant proteins, 92-93 uPA and uPA receptor, 259 VEGF. 289-292 Metastasis. role of-VEGF, 310 MHC, see Major histocompatibility complex &-Microglobulin association with HLA class I heavy chain, 161-162 defects, and HLA class I underexpression, 178- 180 Microsatellite instability in epithelial tumors, 74-77 replication errors leading to, 66 Monoclonal antibodies defining HLA alterations, 171-174 IgM, production in AIDS-SNCCL, 123-124 Monocytes production of SF, 270 VEGF-induced migration, 289 Motility, ECs, SF-induced, 260-261 Mouse lung tumors, genetics, 83-106 Matrigel assay of angiogenesis, 263266 Mutational events, and age-specific incidence of cancer, 5-7 Mutations &m gene, 179-180 germ line, 7-8, 14, 18 induction of SF, 272-274 K w A ~ ,86-91 loss-of-function, 46 multiple recessive, 78 PS?, 91 RER+ phenotype-induced, 68 single-base, and 30-base pair deletions, 217-21 8 somatic. 4-7

N Nasopharyngeal carcinoma EBV+, 199,211-212

tumors, XhoI restriction site loss, 215218 Natural history, AIDS-related lymphomas, 118-120 Natural killer cells, susceptibility, and MHC class I level, 168-169 Neoplasms, AIDS-related, systemic, 116117 Neovascularization corneal, SF-induced dose-dependent, 264-266 persistent, abnormal, 266-267 in psoriasis, 261 Neuroblastoma, germ line mutations, 7-8 Neuroectodermal tumor, primitive, t(l1;22) translocation, 40-43 NF-KB DNA-binding activity, 159-160 p50 and p65 subunits, 177 NHL, see Non-Hodgkin’s lymphoma Nitric oxide, secretion by ECs in response to VEGF, 289 NK, see Natural killer cells N-myc, effect on class I expression, 177178 Non-Hodgkin’s lymphoma AIDS-related B-cell clonal expansions, 118- 120 EBV pathogenic role, 126-130 epidemiology, 1 13- 114 HIV role, 130-131 EBV in, 212 Nuclear proteins, classification, 26-27

0 Oligomerization, Fit receptors, 302 Oncogenes, see also Protooncogenes in AIDS-related lymphomagenesis, 132-137 virus-activated transcription, 10- 12 Oncogenesis, progenitor cell, developmental model, 25-51 Oncogenic viruses, early work, 2-4 Oral hairy leukoplakia EBV+ biopsies, 22 1-224 PCR analysis, 21 7 Orf virus, VEGF-related genes, 296

INDEX

P P53 inactivation in AIDS-SNCCL, 137-138 inhibition of bcl-2 transcription, 272 mutations in glioblastoma multiform, 292 in lung tumors, 91 providing signal for DNA repair, 15-16 Pas1 locus, on chromosome 6, 97-102, 105 Pathogenesis AIDS-related lymphomas, 113- 142 EBV, 244-246 lung adenocarcinoma, KRAS2 activation in, 86-91 PCNSL, see Primary central nervous system lymphomas PCR, see Polymerase chain reaction Peptide binding, to class I groove, 164165 Peptide transporters, selective for peptide length, 163 Pericytes, expression of c-met mRNA, 261-263 Peripheral T-cell lymphomas 11-aa repeats, 2 17 EBV+, 198-199 EBV genotypes, 204-205 Permeability blood vessels, in cancer progression, 282 vascular, stimulation by VEGF, 288 Persistent generalized lymphadenopathy in early AIDS-NHL, 118-120 preceding AIDS-SNCCL, 139-142 presence of EBV in context of, 128129 Phenotypic modulation, in SMCs, 263 Philadelphia chromosome abnormality in CML, 5, 10, 16 break point cloning, 32-33 Phosphorylation, tyrosine, cellular substrates, 303-304 P3HR-1, EBV mutant, 220-224 Placenta growth factor, role in tumor angiogenesis, 295-296 Platelet-derived growth factor, subtype corresponding to VEGF subtype, 294

325

Ploidy, distribution of rearranged chromosomes in relation to, 69-7 1 PML oncogenic domains, in promyelocytic leukemia, 40 Pockets A-F, interaction with peptide residues, 164-165 Polymerase chain reaction analysis of pericyte mRNA, 263 studies of EBV types A and B, 211-213 throat washings positive for, 217 Polymorphisms, see also Restriction fragment length polymorphism allelic, in M. spretw mice, 96-99 BNLFl gene, 214-218 C/D and F/f, 219-220 in class I and class I1 genes, 155-156 EBNA, 236-237 H3-8, 14 in Kras2 second intron, 101-102 TATA, 170-171 Pox viruses, Orf-VEGFs in, 296 Predisposition, inherited, to lung turnorigenesis in mice, 84-95 Primary central nervous system lymphomas, AIDS-related classification, 115-1 16 EBNA-2 and LMP-1 expression, 129130 EBV-positive, 142 histological homogeneity, 117-1 18 Primitive neuroectodermal tumor, t( 11;22) translocation, 40-43 Prognosis, and HLA-negative tumors, 184 Prognostic indicators, SF and tumor angiogenesis, for breast carcinoma, 267 Prognostic marker, BCL-6, 136 Promoter-enhancer elements, adjacent and rearranged, 26-27 Promoter-enhancer regions, Pt family, 308 Proteasomes, p-subunits LMP-1 and LMP-7, 162 Proteins, see also Gap proteins basic helix-loop-helix, 26-33 CBFP, fusion with MYH11, 39 E2A-HLF, 34-36,48 EBNA-2 and LMP-1, 129-130, 141142 EBNAs and LMP-1 and -2,200-204

326

INDEX

Proteins (Continued) EBV-encoded in EBV-transformed cells, 226-23 1 with repeat sequences, 241-243 heat shock, as chaperones, 163-164 injurin, 271 M a , con~plexwith c-MYC, 136 MAX and MAD, 30-31 pRB, 15 RBTNl and RBTN2, 31-32 surfactant, selective expression in lung cells, 92 ZEBRA, 203, 225 Protein synthesis, and disease course alteration, 51 Protooncogenes, virus-activated transcription, 10-12 Psoriasis, neovascularization in, 261

R Rat, cornea assay of angiogenesis, 263266 Kate-limiting events, in carcinogenesis, 16-17 RB1, first human antioncogene, 12-14 KB gene transcript, in lung adenocarcinomas, 92-95 Rearrangements BCL-6, 136 chromosomal, i n tumor subset, 62-64 c-MYC, in AIDS-NHL, 120 continuous occurrence in breast cancer, 70 distribution in non-small-cell lung cancer, 71-73 Receptor-type tyrosine kinase, isolation and structure, 296-304 Recessive cancer genes, discovery, 10- 14 Reciprocal communication, between parenchymal cells and ECs, 306-308 Recombination, EBV, and genotype generation, 222-224 Regions of minimal deletion, in B-cell NHL, 138-139 Remission, ALL, 2 Replication errors. leading to microsatellite instability, 66

RER+ phenotype associated with DNA-diploid tumors without LOH, 66 in breast and endometrial cancers, 7477 induced mutations, 68 Restriction endonuclease, recognition sites, in EBV isolation, 204-205 Restriction fragment length polymorphism association with adenocarcinoma risk, 104 EBV genotypes, 219-220 in study of EBV transmission, 224 retinoblastoma, 14 Ketinoblastoma hereditary, 7-9, 18 loss of RBI function, 60 in study of recessive cancer genes, 1214 Retroviruses, transformation, 3-4 RFLP, see Restriction fragment length polymorphism RNA c-met, short half-life, 273-274 messenger, see Messenger RNA small, EBER-1 and -2, 200-204 RNA tumor viruses, see Retroviruses Rous sarcoma virus, temperature-sensitive mutants, 3-4

S Salivary contact, EBV infection via, 198, 23 1, 244 Sarcomas, see also Cancer; Tumors chimeric transcription factors, 40-43 Scatter factor a-chain and @-chain,257-258 induced by mutations, 272-274 as potential tumor angiogenesis factor, 266-27 1 producer cell and target cell ECs as, 259-261 SMCs as, 261-263 role in angiogenesis, 27 1-275 Segmentation, in Drosophila, 44

INDEX

Serine protease, produced in zymogen form, 258 SF, see Scatter factor SF-inducing factors, regulation of SF production, 270-271 Signal transduction, Flt family, 302-304 Simultaneous carriage, multiple ebnotypes, 232-234 Small noncleaved cell lymphomas, AIDSrelated c-MYC activation, 132-136 disrupted immunosurveillance, 120- 122 IgM mAb production, 123-124 systemic, 116-1 17 SMC, see Smooth muscle cells Smokers, see also Tobacco exposure adenocarcinomas from, KRAS2 mutations, 90 Smooth muscle cells, vascular, as SF producer cell and target cell, 261-263 SNCCL, see Small noncleaved cell lymphomas Somatic mutations, research history, 4-7 Southern blotting, uncultured tumor specimens, 2 10-2 1 1 Susceptibility lung tumor, mapped loci affecting, 101-102 to lung tumorigenesis, 83-85, 98-99 SV40 large T antigen, ransgenic mice expressing, 99-100

T TATA polymorphism, 70-171 T cell receptors, interactions with foreign epitopes, 167-168 T cells cytotoxic, see CTL tumor-infiltrating, 122 Throat washings from apparently healthy adults, EBNA-2 deletions, 221-224 from apparently healthy EBV carriers, 2 13 PCR-positive, 217 Thrombospondin 1, angioinhibitory activity, 271-272

327

Thymidine metabolism, in colorectal cancers, 64-65 Timing, HLA alterations, 174- 175 Tissues normal, HLA class I antigen expression, 157-166 normal adult, mRNA levels j3t-1, 305-308 VEGF, 290 T lymphocytes, see T cells Tobacco exposure, see also Smokers and pattern of lung cancer, 103- 104 Transcription, bcl-2, inhibition by p53, 272 Transcriptional control genes, alteration by chromosomal rearrangements, 2643 Transcription factors master, developmental model, 25-51 NF-KB,DNA-binding activity, 159-1 60 oncogenic, and Drosophilu regulatory proteins, 43-50 Transformation E2A-HLF-mediated, 36-38 by MYC, 30-31 retroviruses, 3-4 Transgenic models, lung tumorigenesis, 99-100 Translocations affecting band 8q24, 132- 134 in ALL, 27-38 T-cell, reverse genetics at break points, 50 tumor-specific, 10-12 Transmission, EBV during bone marrow transplant, 198199, 235-236, 244 existing concepts, 244 via salivary contact, 198, 231, 244 study with RFLP, 224 Tubular formation, ECs, 286-288 Tumor antigens, and HLA, 170- 17 1 Tumor growth, regulation by suppression of VEGF-Flt receptor system, 309310 Tumorigenesis, lung and P a d locus, 105 susceptibility, 83-85 transgenic models, 99-100

328

INDEX

Tumor necrosis factor a,stimulation of SF production, 270-271 Tumors, see also Cancer; Sarcomas angiogenesis macrophage role, 269-270 PDGF role, 295-296 suppression, 309-3 10 EBV+ malignant, 224-226 expression of SF within, 267-268 HLA-negative, and prognosis, 184 human, HLA class I antigens, 155-185 localization, 64 scatter factor, cellular sources, 268-269 subset evolution, 60-61 Tumor suppressor genes, see also Antioncogenes MTSIICDKNP, 94 p53, 137-138 p53, 100, 272-274 R b l , 93-94 Tumor suppressor loci, role in AIDSNHL, 137-139 Tumor type cytogenetic definition, 62-64 and genetic predisposition, 65-67 and pathways of genomic alteration, 67-68 and variability of chromosome alterations, 77-78 Tyrosine kinase, receptor-type, isolation and structure, 296-304

U uPA expression, SF-induced, 26 I mRNA, SF-stimulated, 259

v Iascular endothelial cells, as SF producer cell and target cell, 2.59-26 1

Vascular endothelial growth factor and Flt receptor gene family, 296308 subtypes, 294-295 and VPF, structure and functions, 283296 Vascular endothelial growth factor-Flt receptors regulation of angiogenesis, 281-282 suppression, 309-310 Vascular permeability factor, and VEGF, structure and functions, 283-296 VEGF, see Vascular endothelial growth factor Viruses, see also specific viruses infection, effect on MHC class I expression, 180-182 oncogenic, 2-4 VPF, see Vascular permeability factor

W Wilms’ tumor, two-mutation hypothesis, 8-9 Wound healing application of SF, 274 and keratinocyte expression of VEGF rnRNA, 291

X XhoI restriction site deletion in LMP-I, 241-243 loss in nasopharyngeal carcinoma tumors, 215-218

Z Zinc finger dependent lymphoid cell development, 49-50 in human acute leukemia, 26-38

This Page Intentionally Left Blank